Home

ILOG CPLEX 8.1 Advanced Reference Manual

Contents

1. CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_SOLVE or CPX_CALLBACK_MIP_CUT Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetcallbacklp CPXCENVptr env void cbdata int wherefrom CPXCLPptr lp_p Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLI routines irl gt lt void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback CPXCLPptr lp_p A pointer to a variable of type CPXLPpt r to receive the pointer to the LP problem object which is a MIP Example status CPXgetcallbacklp env cbdata wherefrom amp origlp See Also Examples admipex1 c admipex2 c and admipex3 c in the advanced examples directory 92 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbacknodeinto CPXgetcallbacknodeinfo Usage Mixed Integer Users Only Description The routine CPXgetcallbacknodeinfo is called from within user written callbacks during a MIP optimization and accesses information about nodes When the wherefrom argument is CPX_CALLBACK_MIP_NODE a node with any nodeindex value can be queried When the wherefrom argument is an
2. Example Here is an example of a file containing ordinary constraints and lazy constraints Maximize 34 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL ADDING USER CUTS AND LAZY CONSTRAINTS obj 12 xl 5 x2 15 x3 10 x4 Subject To cl 5 x1 x2 9 x3 12 x4 lt 15 Lazy Constraints 11 2 x1 3 x2 4 x3 x4 lt 10 12 3 x1 2 x2 4 x3 10 x4 lt 8 Bounds 0 lt x1 lt 5 0 lt x2 lt 5 0 lt x3 lt 5 0 lt x4 lt 5 Generals xl x2 x3 x4 End The optional constraint names in LP format for user cuts and lazy constraints are discarded Writing LP and SAV Files When writing LP or SAV format files user cuts and lazy constraints added through their respective add routines or read from LP or SAV format files will be included in the output files In LP format files 4 User cuts will be given names of the form ux where x is an index number starting at 1 Lazy constraints will be given names of 1x where x is again an index number starting at 1 Using the Interactive Optimizer 00q JUIB1ISUOD Y D e Q 3 Q m N lt User cuts and lazy constraints will appear when the command display problem all is issued in the Interactive Optimizer User cuts and lazy constraints can also be added to an existing problem with the add command of the Interactive Optimizer CPLEX Parameters In the Callable Library CPLEX parameters must be set as follo
3. These reductions are controlled through the CPX_PARAM_REDUCE parameter The parameter has four possible settings The default value CPX_PREREDUCE_PRIMALANDDUAL 3 indicates that presolve can rely on primal and dual information With setting CPX_PREREDUCE_DUALONLY 2 presolve only uses dual information with setting CPX_PREREDUCE_PRIMALONLY 1 it only uses primal information and with setting CPX_PREREDUCE_NO_PRIMALORDUAL 0 it uses neither which is equivalent to turning presolve off mv gt O p o gt 72 2 A 0S91d paoueApY Setting the CPX_PARAM_REDUCE parameter has one additional effect on the optimization Normally the presolved problem and the presolved solution are freed at the end of an optimization call However when CPX_PARAM_REDUCE is set to a value other than its default CPLEX assumes that the problem will subsequently be modified and reoptimized It therefore retains the presolved problem and any presolved solution information internally to the LP problem object If the user has set CPX_PARAM_REDUCE and is finished with problem modification he can call CPXfreepresolve to free the presolved problem and reclaim the associated memory The presolved problem is automatically freed when the user calls CPXfreeprob on the original problem We should note that cutting planes in mixed integer programming are handled som
4. Arguments Example See Also 158 CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata A pointer passed from the optimization routine to the user written callback that identifies the problem being optimized The only purpose of this pointer is to pass it to the callback information routines int wherefrom An integer value indicating where in the optimization this function was called It will have the value CPX_CALLBACK_MIP_DELETENODE void cbhandle A pointer to user private data int seqnum The sequence number of the node that is being deleted void handle A pointer to the user private data that was assigned to the node when it was created with one of the callback branching routines CPXbranchcallbackbranchbds CPXbranchcallbackbranchconstraints or CPXbranchcallbackbranchgeneral status CPXsetdeletenodecallbackfunc env mybranchfunc mydata CPX getdeletenodecallbackfunc CPXbranchcallbackbranchbds CPXbranchcallbackbranchconstraints CPXbranchcallbackbranchgeneral Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsetheuristiccallbackfunc CPXsetheuristiccallbackfunc Usage Description Return Value Synopsis Arguments Callback Description Mixed Integer Users Only The routine CPXsetheuristiccallbackfunc sets or modifies the user written
5. Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 95 CPXgetcallbacknodeintfeas CPXgetcallbacknodeintfeas Usage Mixed Integer Users Only Description The routine CPXgetcallbacknodeintfeas is used to get indicators for each variable of whether or not the variable is integer feasible in the node subproblem It can be used in a user written callback during MIP optimization The indicators are from the original problem if CPX_PARAM_MIPCBREDLP is set to CPX_OFF Otherwise they are from the presolved problem This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC Or CPX_CALLBACK_MIP_CUT Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetcallbacknodeintfeas CPXCENVptr env void cbdata int wherefrom int feas int begin int end Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback
6. Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual Example status CPXfreeusercuts env lp ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 77 CPXtftran CPXftran Usage Description Return Value Synopsis Arguments 78 Advanced The routine CPXftran solves By x and puts the answer in the vector x where B is the basis matrix The routine returns a zero on success and a nonzero if an error occurs int CPXftran CPXCENVptr env CPXCLPptr lp double x CPXCENVptr env E x lt The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPL routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double x An array that holds the right hand side vector on input and the solution vector on output The array must be of length at least equal to the number of rows in the LP problem object ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetbasednorms CPXgetbasednorms Usage Advanced Description The routine CPXgetbasednorms works in conjunction with the routine Return Value CPXcopybasednorms CPXgetbasednorms retrieves the resident basis and dual norms from a specified problem o
7. A double that indicates the value of the node estimate for the node to be created with this branch The node estimate is used to select nodes from the branch amp cut tree with certain values of the NodeSel parameter int rent An integer that indicates the number of constraints for the branch 48 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXbranchcallbackbranchconstraints int nzcnt An integer that indicates the number of nonzero constraint coefficients for the branch This specifies the length of the arrays rmatind and rmatval const double rhs An array of length rent containing the right hand side term for each constraint for the branch const char sense An array of length rent containing the sense of each constraint to be added for the branch sense i L Sconstraint sense i E constraint sense i G 2 constraint const int rmatbeg const int rmatind const double rmatval Arrays that describe the constraints for the branch The format is similar to the format used to describe the constraint matrix in the routine CPXaddrows Every row must be stored in sequential locations in this array from position rmatbeg i to rmatbeg i 1 1 or from rmatbeg i tonzcnt 1 if i rcnt 1 Each entry rmatind i indicates the column index of the corresponding coefficient rmatval i All rows must be contiguous and rmatbeg 0 must be 0 void userhandle A pointer to user private data that should
8. The address of a variable to hold the user s private pointer CPXgetnodecallbackfunc env current_callback amp current_handle The example admipexl c CPXsetnodecallbackfunc CPXgetcallbacknodeinfo Advanced MIP Control Interface CPLEX 8 1 ADVANCED REFERENCE MANUAL 125 CPXgetobjoffset CPXgetobjoffset Usage Advanced Description The routine CPXgetobjoffset returns the objective offset between the original problem and the presolved problem Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetobjoffset CPXCENVptr env CPXCLPptr redlp double objoffset_p Arguments CPXCENVptr env E x lt The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPL routines CPXCLPptr redlp A pointer to a reduced CPLEX LP problem object as returned by CPXgetredlp double objoffset_p A pointer to a variable of type double to hold the objective offset value 126 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetpnorms CPXgetpnorms Usage Advanced Description The routine CPXgetpnorms returns the norms from the primal steepest edge Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetpnorms CPXCENVptr env CPXCLPptr lp double cnorm double rnorm int len_p Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one
9. Thus proof_p in some sense denotes the degree of infeasibility The routine returns a zero on success and a nonzero if an error occurs int CPXdualfarkas CPXCENVptr env CPXCLPptr lp double y double proof_p ILOG ENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLE routines CPLEX 8 1 ADVANCED REFERENCE MANUAL 73 CPXdualfarkas CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double y An array of doubles of length at least equal to the number of rows in the problem double proof_p A pointer to a double The parameter proof_p is allowed to have the value NULL 74 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXfreelazyconstraints CPXfreelazyconstraints Usage Description Return Value Synopsis Arguments Example Mixed Integer Users Only The routine CPXfreelazyconstraints is used to clear the list of lazy constraints that have been previously specified through calls to CPXaddlazyconstraints The routine returns a zero on success and a nonzero if an error occurs int CPXfreelazyconstraints CPXCENVptr env CPXLPptr 1p CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference
10. ILOG CPLEX 8 1 Advanced Reference Manual December 2002 Copyright 2002 by ILOG This document and the software described in this document are the property of ILOG and are protected as ILOG trade secrets They are furnished under a license or non disclosure agreement and may be used or copied only within the terms of such license or non disclosure agreement No part of this work may be reproduced or disseminated in any form or by any means without the prior written permission of ILOG S A Printed in France Table of Contents Preface About This Manual 222222000 cece see eee eee rr eee ee eee end 9 Designating and Modifying User Written Callbacks in MIPS 10 Accessing Current User Written Callbacks in MIPS oooooooommmmmm o 10 Specifying Actions with User Written Callbacks with Routines 10 Querying from Within User Written Callbacks in MIPS o oooooooooo 10 Detecting Round Off Errors in a Solution 1 2 0 0 00 c cece eee 11 Using the Bass oia end edie ead eared ween ca aed a 11 Accessing and Manipulating Norms 00 eee eee eee eee eee 12 Accessing Information About the Current Problem 200 eee eee eee eeee 12 Changing the Basis by Pivoting 0 eee e eee eee eee eens 12 Using User Cuts and Lazy Constraints ooooocoocoonnncnr eee 12 Advanced Presolve Routines 00 cee eee eee eee eee eee eee 12 Miscellan
11. Usage Advanced Description The routine CPXpreaddrows is used to add rows to an LP problem object and its associated presolved LP problem object Note that the CPLEX parameter CPX_PARAM_REDUCE must be set to CPX_PREREDUCE_PRIMALONLY 1 or CPX_PREREDUCE_NOPRIMALORDUAL 0 at the time of the presolve in order to add rows and preserve the presolved problem This routine should be used in place of CPXaddrows when it is desired to preserve the presolved problem Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXpreaddrows CPXENVptr env CPXLPptr lp int rent int nzcnt double rhs char sense int rmatbeg int rmatind double rmatval char rowname Arguments The arguments of CPXpreaddrows are the same as those of CPXaddrows with the exception that new columns may not be added so there are no ccnt and colname arguments The new rows are added to both the original LP problem object and the associated presolved LP problem object Examples status CPXpreaddrows env lp rent nzcnt rhs sense rmatbeg rmatind rmatval newrowname See Also Example adpreex1 c in the advanced examples directory 144 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXprechgobj CPXprechgobj Usage Advanced Description The routine CPXprechgob3 is used to change the objective function coefficients of an LP problem object an
12. can be used in the branch incumbent and heuristic callbacks to obtain information on the current node Advanced Presolve Functions for QP Many of the advanced presolve routines have been extended for QP such as CPXcrushx and CPXuncrushx However two new routines are needed for transforming dual values between the original and presolved QP problems CPXqpuncrushpi and CPXqpdjfrompi CPLEX 7 1 Advanced Features Release Notes CPLEX 7 1 offers a number of new routines and extensions to existing routines both in the Callable Library and Concert Technology Library Conversion Notes The following routines have been modified The branch callback function has been changed so that additional information is available to the user and so that users may branch from integer feasible nodes Previously the branches to be created in a branch callback were specified by filling the arguments to the callback function in 7 1 a new function CPXbranchcallbackbranchbds is called to specify the branches The node sequence number assigned to the node created by the specified branch is returned See the sections for e CPXsetbranchcallbackfunc e CPXbranchcallbackbranchbds 4 In the Concert Technology Library IloCplex importModel methods are extended with extra parameters for user cuts and lazy constraints ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 193 CPLEX 7 1 ADVANCED FEATURES RELEASE NOTES New Features In
13. int method Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int method Example An integer specifying the optimization algorithm to be used to solve the problem after the presolve is completed Some presolve reductions are specific to an optimization algorithm so specifying the algorithm ensures that the problem is presolved for that algorithm and that presolve does not have to be re done when that optimization routine is called Possible values are CPX_ALG_NONE CPX_ALG_PRIMAL CPX_ALG_DUAL and CPX_ALG_BARRIER for LP CPX_ALG_NONE should be used for MIP status CPXpresolve env lp CPX_ALG_DUAL 146 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXaqpdjfrompi CPXqpdjfrompi Usage Description Return Value Synopsis Arguments Example Advanced The routine CPXapd3frompi computes an array of reduced costs from an array of dual values for a QP The routine returns a zero on success and a nonzero if an error occurs int CPXqpdjfrompi CPXCENVptr env CPXCLPptr lp const double pi const double x double dj CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object
14. 118 136 141 dual values using to compute reduced costs 72 F fathoming nodes 152 feasibility checking for 160 files format for solution 170 node 166 solution 170 fixed variable changing bounds 175 definition 143 pivoting out of basis 143 freePresolve member function IloCplex class 183 G getDiverging member function IloCplex class 184 getRay member function IloCplex class 184 H heuristic callback 27 120 159 llo IloCplex class 182 addLazyConstraint member function 182 addLazyConstraints member function 182 addUserCut member function 182 addUserCuts member function 183 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 203 INDEX basicPresolve member function 183 clearLazyConstraints member function 183 clearUserCuts member function 183 freePresolve member function 183 getDiverging member function 184 getRay member function 184 importModel member function 184 presolve member function 185 IloCplex LazyConstraintCallbackI class 186 defining with macro 188 IloCplex UserCutCallbackI class 187 defining with macro 189 LOLAZYCONSTRAINTCALLBACK macro 188 OUSERCUTCALLBACK macro 189 H H I continued importModel member function TloCplex class 184 incumbent callback before replacing 123 162 incumbent callback 30 162 incumbent values accessing from user written callback 90 indexing bounds on variables 175 columns and dual steepest edge norms 58 118 cuts 154 divergent column 121
15. Aj The routine returns a zero on success and a nonzero if an error occurs int CPXbinvacol CPXCENVptr env CPXCLPptr lp int Jy double x Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int j An integer that indicates the index of the column to be computed double x 42 ILOG An array containing the solution of Bx Aj The array must be of length at least equal to the number of rows in the problem CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXbinvarow CPXbinvarow Usage Advanced Description The routine CPXbinvarow computes the if row of BA In other words it computes Return Value Synopsis Arguments the 2 row of the tableau The routine returns a zero on success and a nonzero if an error occurs int CPXbinvarow CPXCENVptr env CPXCLPptr lp int i double z CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int i An integer that indicates the index of the row to be computed double z An array containing the i row of B A The array must be of length at least equal to the number
16. CPXdualfarkas CPXdualfarkas Usage Description Return Value Synopsis Arguments CPXCI Advanced The routine CPXdualfarkas assumes that there is a resident solution as produced by a call to CPXdualopt and that the status of this solution as returned by CPXgetstat is CPX_UNBOUNDED The values returned in the array y have the following interpretation For the i constraint if that constraint is a less than or equal to constraint y i lt O holds if that constraint is a greater than or equal to constraint y i gt O holds Thus where b is the right hand side vector for the given linear program A is the constraint matrix and x denotes the vector of variables y may be used to derive the following valid inequality YA x gt y b Here y is being interpreted as a column vector and y denotes the transpose of y The real point of computing y is the following Suppose we define a vector z of dimension equal to the dimension of x and having the following value for entries zj uj Where yA gt 0 and T zj lj where y A lt 0 where A denotes the column of A corresponding to x u the given upper bound and J is the specified lower bound z is arbitrary if YA 0 Then y and z will y b y A z gt 0 This last inequality contradicts the validity of yA x2 yb and hence shows that on Xj satisfy the given linear program is infeasible The quantity proof_p is set equal to yb yA Zi
17. amp current_callback amp current_handle ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetbranchcallbackfunc See Also CPXsetbranchcallbackfunc Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 83 CPXgetcallbackctype CPXgetcallbackctype Usage Mixed Integer Users Only Description The routine CPXgetcallbackctype is used to get the ctypes for the MIP problem from within a user written callback during MIP optimization The values are from the original problem if CPX_PARAM_MIPCBREDLP is set to CPX_OFF otherwise they are from the presolved problem This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_SOLVE or CPX_CALLBACK_MIP_CUT Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetcallbackctype CPCXENVptr env void cbdata int wherefrom char xctype int begin int end Arguments CPCXENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called
18. as returned by CPXcreateprob documented in the CPLEX Reference Manual const double pi An array that contains dual solution pi values for a problem as returned by such routines as CPXqpuncrushpi and CPXcrushpi The length of the array must at least equal the number of rows in the LP problem object const double x An array that contains primal solution x values for a problem as returned by such routines as CPXuncrushx and CPXcrushx The length of the array must at least equal the number of columns in the LP problem object double dj An array to receive the reduced cost values computed from the pi values for the problem object The length of the array must at least equal the number of columns in the problem object status CPXqpdjfrompi env lp origpi reducepi ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 147 CPXapdjfrompi 148 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXqpuncrushpi CPXqpuncrushpi Usage Description Return Value Synopsis Arguments Example Advanced The routine CPXqpuncrushpi uncrushes a dual solution for the presolved problem to a dual solution for the original problem if the original problem is a QP The routine returns a zero on success and a nonzero if an error occurs int CPXqpuncrushpi CPXENVptr env CPXLPptr lp double pi const double prepi const double x CPXENVptr env The pointer to the CPLEX environment as returned by
19. divergent constraint 121 divergent row 121 divergent variable 121 nodes 94 166 position of basic variable in basis header 122 rows and dual steepest edge norms 58 118 special ordered sets SOS 114 variables 81 infeasibilities callbacks and 120 159 integer and cuts 154 number of 94 sum of 94 integer solution 120 159 173 K kappa estimate 124 L L_infinity norm 54 55 lazy constraint 28 adding 182 callback 186 deleting 183 pool 33 to 36 lazy constraints clearing 75 LazyConstraintCallbackI nested class of 110C plex 186 linear formula crushing original to presolved 63 uncrushing presolved to original 177 local cut adding 70 lower bound values accessing for subproblem from user written callback 98 accessing from user written callback 86 macro ILOLAZYCONSTRAINTCALLBACK 188 ILOUSERCUTCALLBACK 189 MIP optimizer adding cuts to subproblems 67 70 callbacks after node selection 125 165 callbacks for cuts 116 117 154 157 callbacks on branch selection 82 150 callbacks on heuristics 120 159 callbacks on incumbent 123 162 callbacks optimizing subproblems 135 167 cutting planes and 141 information about callbacks 93 100 112 information about SOS in callbacks 114 strong branching 173 modifying bounds of subproblem 159 204 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL branch callback 150 branch selection 150 173 cut callback 154 heuristic callback 159 incumbent 162 node callback 165
20. 1 ADVANCED REFERENCE MANUAL CPXgetcallbacknodeub int end An integer indicating the end of the range of upper bound values to be returned Example status CPXgetcallbacknodeub env cbdata wherefrom ub 0 cols 1 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 105 CPXgetcallbacknodex CPXgetcallbacknodex Usage Description Return Value Synopsis Arguments 106 Mixed Integer Users Only The routine CPXgetcallbacknodex is used to get the primal variable x values for the subproblem at the current node during MIP optimization from within a user written callback The values are from the original problem if CPX_PARAM_MIPCBREDLP is set to CPX_OFF otherwise they are from the presolved problem This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC Or CPX_CALLBACK_MIP_CUT The routine returns a zero on success and a nonzero if an error occurs int CPXgetcallbacknodex CPXCENVptr env void cbdata int wherefrom double x int begin int end CPXCENVptr env sa pS The pointer to the CPLEX environment as returned by one of the CPXopenCPLI routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefro
21. A pointer to a double to contain the value of the linear formula corresponding to variables that have been removed in the presolved problem ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 63 CPXcrushtorm Example int pind double pval The linear formula in terms of the presolved problem Each entry pind i indicates the column index in the presolved problem of the corresponding coefficient pval 1 The arrays pind and pval must be of length at least the number of columns in the presolved LP problem object status CPXcrushform env lp len ind val amp plen amp poffset pind pval ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXcrushpi CPXcrushpi Usage Advanced Description The routine CPXcrushpi crushes a dual solution for the original problem to a dual solution for the presolved problem Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXcrushpi CPXCENVptr env CPXCLPptr lp const double pi double prepi Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual const double pi An array that contains dual solution pi values for the original problem as returned by routines such as CPXgetpi or CPXsolution The array must be of length at
22. A pointer to an integer that indicates the number of entries in both norm and head The value assigned to the pointer len_p is needed by the routine CPXcopydnorms CPXcopydnorms ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetExactkappa CPXgetExactkappa Usage Advanced Description The routine CPXgetExactkappa computes and returns the condition number kappa Return Value If successful this routine returns the condition number kappa Synopsis double CPXgetExactkappa CPXCENVptr env CPXCLPptr lp Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual See Also CPXgetkappa ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 119 CPXgetheuristiccallbackftunc CPXgetheuristiccallbackfunc Usage Description Return Value Synopsis Arguments Example See Also 120 Mixed Integer Users Only The routine CPXgetheuristiccallbackfunc accesses the user written callback to be called by ILOG CPLEX during MIP optimization after the subproblem has been solved to optimality That callback is not called when the subproblem is infeasible or cut off The callback supplies ILOG CPLEX with heuristically derived integer solutions For documentation of callback arguments see the routine CPXsetheuristicca
23. Arguments CPXENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPL routines El gt CPXLPptr lp A pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int idiv_p A pointer to an integer indexing the row of a diverging variable If one of the ILOG CPLEX simplex optimizers has concluded that the LP problem object is unbounded and if the diverging variable is a slack or ranged variable CPXget ijdiv returns the index of the corresponding row in idiv_p Otherwise idiv_p is set to 1 int jdiv_p A pointer to an integer indexing the row of a diverging variable If one of the ILOG CPLEX simplex optimizers has concluded that the LP problem object is unbounded and if the diverging variable is a normal structural variable CPXgetijdiv sets jdiv_p to the index of that variable Otherwise jdiv_pis set to 1 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 121 CPXgetijrow CPXgetijrow Usage Description Return Value Synopsis Arguments CPXC Advanced The routine CPXget ijrow returns the index of a specific basic variable as its position in the basis header If the specified row indexes a constraint that is not basic or if the specified column indexes a variable that is not basic CPXgetijrow returns an error code and sets the value of its argument row_p to 1 An error is also returned if both row a
24. CPLEX calls the node callback after selecting the next node to explore The user can choose another node by setting the argument values of the callback Return Value The callback returns a zero on success and a nonzero if an error occurs Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPL routines ira x lt ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 165 CPXsetnodecallbackfunc void cbdata A pointer passed from the optimization routine to the user written callback that identifies the problem being optimized The only purpose of this pointer is to pass it to the callback information routines int wherefrom An integer value indicating where in the optimization this function was called It has the value CPX_CALLBACK_MIP_NODE void cbhandle A pointer to user private data int nodeindex_p A pointer to an integer that indicates the node number of the user selected node The node selected by ILOG CPLEX is node number 0 zero Other nodes are numbered relative to their position in the tree and this number changes with each tree operation The unchanging identifier for a node is its sequence number To access the sequence number of a node use the ILOG CPLEX Callable Library routine CPXgetcallbacknodeinfo An error results if a user attempts to select a node that has been moved to a node file See the CPLEX User s Manual for more information about no
25. Constant Value Branch CPX_TYPE_VAR 0 variable branch CPX_TYPE_SOS1 TI SOS1 branch CPX_TYPE_SOS2 rzi SOS2 branch CPX_TYPE_USER Xx user defined int sos An integer that indicates the special ordered set SOS used for this branch A value of 1 indicates that this branch is not an SOS type branch int nodecnt An integer that indicates the number of nodes CPLEX will create from this branch Possible values are 0 zero 1 and 2 If the argument is 0 the node will be fathomed unless user specified branches are made that is no child nodes are created and the node itself is discarded int bdent An integer that indicates the number of bound changes defined in the arrays indices lu and bd that define the CPLEX selected branch double nodeest An array with nodecnt entries that contains estimates of the integer objective function value that will be attained from the created node int nodebeg An array with nodecnt entries The i entry is the index into the arrays indices lu and bd of the first bound changed for the i node int indices Together with 1u and ba this array defines the bound changes for each of the created nodes The entry indices i is the index for the variable char lu Together with indices and bd this array defines the bound changes for each of the created nodes The entry lu i is one of the three possible values indicating which bound to change L for lower bound U for upper bound or
26. FEATURES RELEASE NOTES CPLEX 6 6 to 7 0 Advanced Features Release Notes Conversion Notes The following routines replace the routine CPXgetvarcallbackinfo 9 06 8 4 9 9 9 9 8 44 4 2 e e E e Cc E C Cc E E e E e PXge PXge PXge PXge PXge PXge PXge PXge PXge PXge PXge PXge PXge tcallbacknodex tcallbacknodeobjval tcallbackctype tcallbackorder tcallbackpseudocosts tcallbackincumbent tcallbacknodeintfeas tcallbackgloballb tcallbackglobalub tcallbacknodelb tcallbacknodeub tcallbacknodestat tcallbacklp The following routines have been renamed ec ec ec PXge PXge PXge tnodecallbackinfo has been renamed CPXgetcallbacknodeinfo tsoscallbackinfo has been renamed CPXgetcallbacksosinfo tsubcallbackinfo has been renamed CPXgetcallbacknodelp New Features Advanced Presolve An Advanced Presolve Interface has been added It includes features for controlling presolve access to the presolve problem and accessing original variables from within MIP control callbacks CPXpivot Routine A routine to specify a simplex pivot CPXpivot has been added CPXgetray Routine A routine to return an unbounded direction vector for primal unbounded or dual infeasible problems has been added CPXget ray 196 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL List of Tables Table 1 Information Requested for a User Written Node Callback 94 Table 2 Br
27. Manual status CPXfreelazyconstraints env 1p ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 75 CPXfreepresolve CPXfreepresolve Usage Advanced Description The routine CPXfreepresolve frees the presolved problem from the LP problem object Under the default setting of CPX_PARAM_REDUCE the presolved problem is freed when an optimal solution is found It is not freed when CPX_PARAM_REDUCE is set to CPX_PREREDUCE_PRIMALONLY 1 or CPX_PREREDUCE_DUALONLY 2 so the routine CPXfreepresolve can be used to free it manually Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXfreepresolve CPXCENVptr env CPXLPptr 1p Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual Example status CPXfreepresolv env lp 76 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXfreeusercuts CPXfreeusercuts Usage Mixed Integer Users Only Description The routine CPXfreeusercuts is used to clear the list of user cuts that have been previously specified through calls to CPXaddusercuts Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXfreeusercuts CPXCENVptr env CPXLPptr lp
28. Path TloCplex CallbackI TloCplex MIPCallbackI TloCplex ControlCallbackI TloCplex CutCallbackI IloCplex UserCutCallbackI Description An instance of the class IloCplex UserCutCallbackI represents a user written callback in an application that uses an instance of I1oCplex to solve a MIP while generating user cuts to tighten the LP relaxation IloCplex calls the user written callback after solving each node LP exactly like I10Cplex CutCallbackI The only difference to TloCplex CutCallbackI is that constraints added in a UserCutCallbackI must be real cuts in the sense that omitting them does not affect the feasible region of the model under consideration Include File lt ilcplex ilocplex h gt Definition File lt ilcplex ilocplexi h gt Synopsis class UserCutCallbackI public CutCallbackI protected UserCutCallbackI Y ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 187 ILOLAZYCONSTRAINTCALLBACK Category Synopsis Description Include File Definition File See Also Macro ILOLAZYCONSTRAINTCALLBACKO name ILOLAZYCONSTRAINTCALLBACK1 name typel x1 ILOLAZYCONSTRAINTCALLBACK2 name typel xl type2 x2 ILOLAZYCONSTRAINTCALLBACK3 name typel xl type2 x2 type3 x3 ILOLAZYCONSTRAINTCALLBACK4 name typel xl type2 x2 type3 x3 type4 x4 ILOLAZYCONSTRAINTCALLBACK5 name typel xl type2 x2 type3 x3 type4 x4 type5 x5 ILOLAZYCONSTRAINTCALLB
29. The parameter must be the value of wherefrom passed to the user written callback char xctype An array where the ctype values for the MIP problem will be returned The array must be of length at least end begin 1 If successful xctype 0 through xctype end begin contain the variable types int begin An integer indicating the beginning of the range of ctype values to be returned 84 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbackctype int end An integer indicating the end of the range of ctype values to be returned Example status CPXgetcallbackctype env cbdata wherefrom prectype 0 precols 1 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 85 CPXgetcallbackgloballb CPXgetcallbackgloballb Usage Description Return Value Synopsis Arguments 86 Mixed Integer Users Only The routine CPXgetcallbackgloballb is used to get the best known global lower bound values during MIP optimization from within a user written callback The values are from the original problem if cPX_PARAM_MIPCBREDLP is set to CPX_OFF otherwise they are from the presolved problem This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_SOLVE CPX_CALLBACK_MIP_CUT or The routine returns a zero on success
30. This parameter must be the value of wherefrom passed to the user written callback double nodeest A double that indicates the value of the node estimate for the node to be created with this branch The node estimate is used to select nodes from the branch amp cut tree with certain values of the NodeSel parameter int cnt An integer that indicates the number of bound changes that are specified in the arrays indices lu and bd 46 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXbranchcallbackbranchbds int indices Together with 1u and ba this array defines the bound changes for the branch The entry indices i is the index for the variable const char lu Together with indices and ba this array defines the bound changes for each of the created nodes The entry 1u i is one of the three possible values indicating which bound to change L for lower bound U for upper bound or B for both bounds const int bd Together with indices and lu this array defines the bound changes for each of the created nodes The entry bd i indicates the new value of the bound void userhandle A pointer to user private data that should be associated with the node created by this branch May be NULL int seqnum_p A pointer to an integer that on return will contain the sequence number that CPLEX has assigned to the node created from this branch The sequence number may be used to select this node in later calls to the node callbac
31. a list of inequality rows into the basis If equality rows appear among those specified on the list they are ignored Motivation In the implementation of cutting plane algorithms for integer programming it is occasionally desirable to delete some of the added constraints that is cutting planes when they no longer appear to be useful If the slack on some such constraint that is row is not in the resident basis the deletion of that row may destroy the quality of the basis Pivoting the slack in before the deletion avoids that difficulty Dual Steepest Edge Norms If one of the dual steepest edge algorithms is in use when this routine is called the corresponding norms are automatically updated as part of the pivot Primal steepest edge norms are not automatically updated in this way because in general the deletion of rows invalidates those norms The routine returns a zero on success and a nonzero if an error occurs int CPXpivotin CPXCENVptr env CPXLPptr lp const int rlist int rlen ENVptr env E gt x lt The pointer to the ILOG CPLEX environment as returned by one of the CPXopencPLI routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual const int rlist ILOG An array of length r1en containing distinct row indices of slack variables that are not basic in the current solution If r1ist contains negative entr
32. and a nonzero if an error occurs int CPXgetcallbackgloballb CPXCENVptr env void cbdata int wherefrom double lb int begin int end CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback double lb An array to receive the values of the global lower bound values This array must be of length at least end begin 1 If successful 1b 0 through 1b end begin contain the global lower bound values int begin An integer indicating the beginning of the range of lower bound values to be returned ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbackgloballb int end An integer indicating the end of the range of lower bound values to be returned Example status CPXgetcallbackgloballb env cbdata wherefrom glb 0 cols 1 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 87 CPXgetcallbackglobalub CPXgetcallbackglobalub Usage Description Return Value Synopsis Arguments 88 Mixed Integer Users Only The routine CPXgetcallbackglobalub is used to get the best known global upper bound values during MIP opti
33. cutoff and is integer infeasible The callback routine adds globally valid cuts to the LP subproblem The cut may be for the original problem if the parameter CPX_PARAM MIPCBREDLP was set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise the cut is for the presolved problem Within the user written cut callback the routine CPXgetcallbacknodelp and other query routines from the Callable Library access information about the subproblem The routines CPXgetcallbacknodeintfeas and CPXgetcallbacksosinfo examines the status of integer entities The routine CPXcutcallbackadad adds cuts to the problem Cuts added to the problem are first put into a cut pool so they are not present in the subproblem LP until after the user written cut callback is finished Any cuts that are duplicates of cuts already in the subproblem are not added to the subproblem Cuts that are added remain part of all subsequent subproblems there is no cut deletion If cuts have been added the subproblem is re solved and evaluated and if the LP solution is still integer infeasible and not cut off the cut callback is called again If the problem has names user added cuts have names of the form Xnumber where number is a sequence number among all cuts generated The parameter CPX_PARAM_REDUCE must be set to CPX_PREREDUCE_PRIMALONLY 1 or CPX_PREREDUCE_NOPRIMALORDUAL 0 if the constraints to be added in the callback are lazy co
34. double that must be of length at least goodien For a given j goodlist il upratio i has the following meaning Let Xj be the name of the basic variable with index j and suppose that x is fixed to some value ft gt Then in a subsequent call to CPXdualopt the leaving variable in the first iteration of this call is uniquely determined It must be x There are then two possibilities Either an entering variable is determined or it is concluded in the first iteration that the changed model is dual unbounded primal infeasible In the latter case upratio i is set equal to a large positive value this number is system dependent but is usually 0 75 In the former case where r is the value of the objective function after this one iteration upratio i is determined by Irl t t F upratio i ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 139 CPXpivot CPXpivot Usage Description Return Value Synopsis Arguments Example 140 Advanced The routine CPXpivot performs a basis change where variable jenter replaces variable jleave in the basis It is invalid to pass a basic variable for jenter Also no nonbasic variable may be specified for jleave except for jenter jleave when the variable has both finite upper and lower bounds In that case the variable is moved from the current to the other bound No shifting or perturbation is performed This routine returns a zero on success and a nonzero i
35. get the L norm for the scaled problem set the parameter scalrimtype 1 If successful the routine returns the L norm of Ax b where x is the resident solution If no such solution exists 1 0 is returned double CPXcheckax CPXCENVptr env CPXCLPptr lp int imax_p int scalrimtype ENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int imax_p A pointer to the index of the row with the maximum absolute value in Ax b If no solution exists imax_p is set to 1 int scalrimtype 54 ILOG An integer that indicates the type of scaling to be applied to the returned L norm When this parameter is equal to 0 zero the returned L norm will be unscaled Otherwise the L norm has the same scaling as that applied to the problem currently in memory CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXcheckpib CPXcheckpib Usage Description Return Value Synopsis Arguments Advanced The routine CPXcheckpib finds the L norm of cg B where 7 represents dual solution values and B represents the basis That is this routine checks for numerical roundoff error in the computation of T by putting 7 into the equation that defines it and then returning the value of the maximum deviation from zero of the elements of the resu
36. if CPX_PARAM_MIPCBREDLP is set to CPX_OFF otherwise they are from the presolved problem This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_SOLVE or CPX_CALLBACK_MIP_CUT Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetcallbacknodelb CPXENVptr env void cbdata int wherefrom double lb int begin int end Arguments CPXENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback double lb An array to receive the values of the lower bound values This array must be of length at least end begin 1 If successful 19 0 through 1b end begin contain the lower bound values for the current subproblem int begin An integer indicating the beginning of the range of lower bounds to be returned 98 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbacknodelb int end An integer indicating th
37. int feas An array to receive an indicator of feasibility for the node subproblem This array must be of length at least end begin 1 If successful feas 0 through feas end begin will contain the indicators Values for feas j CPX_INTEGER_FEASIBLE 0 variable j begin is integer valued CPX_INTEGER_INFEASIBLE 1 variable j begin is not integer valued 96 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbacknodeintfeas CPX_IMPLIED_INTEGER_FEASIBLE 2 variable j begin may have a fractional value in the current solution but it will take on an integer value when all integer variables still in the problem have integer values lt should not be branched upon int begin An integer indicating the beginning of the range of feasibility indicators to be returned int end An integer indicating the end of the range of feasibility indicators to be returned Example status CPXgetcallbacknodeintfeas env cbdata wherefrom feas 0 cols 1 See Also Examples admipex1 c and admipex2 c in the advanced examples directory ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 97 CPXgetcallbacknodelb CPXgetcallbacknodelb Usage Mixed Integer Users Only Description The routine CPXgetcallbacknodelb is used to get the lower bound values for the subproblem at the current node during MIP optimization from within a user written callback The values are from the original problem
38. least the number of rows in the LP problem object double prepi An array to receive dual values corresponding to the presolved problem The array must be of length at least the number of rows in the presolved problem object Example status CPXcrushpi env lp origpi reducepi ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 65 CPXcrushx CPXcrushx Usage Advanced Description The routine CPXcrushx crushes a solution for the original problem to a solution for the presolved problem Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXcrushx CPXCENVptr env CPXCLPptr lp const double x double prex Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual const double x An array that contains primal solution x values for the original problem as returned by routines such as CPXgetx or CPXsolution The array must be of length at least the number of columns in the problem object double prex An array to receive the primal values corresponding to the presolved problem The array must be of length at least the number of columns in the presolved problem object Example status CPXcrushx env lp origx reducex See Also Example admipex6 c in the advanced ex
39. lt ilcplex ilocplex h gt lt ilcplex ilocplex h gt TloCplex LazyConstraintCallbackI 188 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL ILOUSERCUTCALLBACK Category Synopsis Description Include File Definition File See Also Macro ILOUSERCUTCALLBACKO name ILOUSERCUTCALLBACK1 name typel x1 ILOUSERCUTCALLBACK2 name typel x1 type2 x2 ILOUSERCUTCALLBACK3 name typel x1 type2 x2 type3 x3 ILOUSERCUTCALLBACK4 name typel xl type2 x2 type3 x3 type4 x4 ILOUSERCUTCALLBACK5 name typel xl type2 x2 type3 x3 type4 x4 type5 x5 ILOUSERCUTCALLBACK6 name typel xl type2 x2 type3 x3 type4 x4 type5 x5 type6 x6 ILOUSERCUTCALLBACK7 name typel xl type2 x2 type3 x3 type4 x4 type5 x5 type6 x6 type 7 x7 This macro creates two things an implementation class for a user defined user cut callback named name and a function named name that creates an instance of this class and returns an IloCplex Callback handle for it This function needs to be called with an environment as first parameter followed by the n parameters specified at the macro execution in order to create a callback The callback can then be used passing it the use method of an I10Cplex object The class name that is created by the macro includes the implementation of method makeClone as required f
40. of columns in the problem ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 43 CPXbinvcol CPXbinvcol Usage Advanced Description The routine CPXbinvcol computes the j column of the basis inverse Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXbinvcol CPXCENVptr env CPXCLPptr lp int j double x Arguments CPXCENVptr env E x lt The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int j An integer that indicates the index of the column of the basis inverse to be computed double x An atray containing the j column of B The array must be of length at least equal to the number of rows in the problem 44 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXbinvrow CPXbinvrow Usage Advanced Description The routine CPXbinvrow computes the i row of the basis inverse Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXbinvrow CPXCENVptr env CPXCLPptr lp int i double y Arguments CPXCENVptr env E x lt The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines CPXCLPptr lp A pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the
41. one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double pi An array to receive dual solution pi values for the original problem as computed from the dual values of the presolved problem object The length of the array must at least equal the number of rows in the LP problem object const double prepi An array that contains dual solution pi values for the presolved problem as returned by such routines as CPXgetpi and CPXsolution when applied to the presolved problem object The length of the array must at least equal the number of rows in the presolved problem object const double x An array that contains primal solution x values for a problem as returned by such routines as CPXuncrushx and CPXcrushx The length of the array must at least equal the number of columns in the LP problem object status CPXqpuncrushpi env lp pi prepi x ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 149 CPXsetbranchcallbackfunc CPXsetbranchcallbackfunc Usage Mixed Integer Users Only Description The routine CPXsetbranchcallbackfunc sets and modifies the user written callback routine to be called after a branch has been selected but before the branch is carried out during MIP optimization In the callback routine the CPLEX selected branch can be changed to a user selected branch Return Value The r
42. the presolved problem The callback returns a zero on success and a nonzero if an error occurs CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPL routines E gt void cbdata A pointer passed from the optimization routine to the user written callback that identifies the problem being optimized The only purpose of this pointer is to pass it to the callback information routines int wherefrom An integer value indicating where in the optimization this function was called It will have the value CPX_CALLBACK_MIP_BRANCH void cbhandle A pointer to user private data double objval A variable that contains the objective value of the integer solution double x An array that contains primal solution values for the integer solution int isfeas_p A pointer to an integer variable that indicates whether or not CPLEX should use the integer solution specified in x to replace the current incumbent A nonzero value indicates that the incumbent should be replaced by x a zero value indicates that it should not int useraction_p A pointer to an integer to contain the indicator for the action to be taken on completion of the user callback Table 9 summarizes possible values Table 9 Actions to be Taken after a User Written Incumbent Callback Value Symbolic Constant Action 0 CPX_CALLBACK_DEFAULT Proceed with optimization 1 CPX_CALLBACK_FAIL Exit optimiza
43. these values CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_BRANCH or CPX_CALLBACK_MIP_CUT The information returned is for the original problem if the parameter CPX_PARAM MIPCBREDLP is set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise it is for the presolved problem The routine returns a zero on success and a nonzero if an error occurs If the return value is nonzero the requested value may not be available int CPXgetcallbacksosinfo CPXCENVptr env void cbdata int wherefrom int sosindex int member int whichinfo void result_p CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating where the user written callback was called from This parameter must be the value of wnerefrom passed to the user written callback int sosindex The index of the special ordered set SOS for which information is requested SOSs are indexed from 0 zero to numsets 1 where numset s is the result of calling this routine with a whichinfo value of CPX_CALLBACK_INFO_SOS_NUM int member The index of the member of the SOS for which information is requested ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbacksosinto int whic
44. to CPXprimopt with a setting of CPX_PARAM_PGRADIENT of 2 or 3 as the initial values for the primal steepest edge norms of the first len columns in the LP problem object The array must be of length at least equal to the value of the argument len double rnorm An array containing values to be used in a subsequent call to CPXprimopt with a setting of CPX_PARAM_PGRADIENT of 2 or 3 as the initial values for the primal steepest edge norms of the slacks and ranged variables that are nonbasic The array must be of length at least equal to the number of rows in the LP problem object int len An integer that indicates the number of entries in the array cnorm See Also CPXcopydnorms CPXgetpnorms ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 61 CPXcopyprotected CPXcopyprotected Usage Description Return Value Synopsis Arguments Example 62 Advanced The routine CPXcopyprotected is used to specify a set of variables that should not be substituted out of the problem If presolve can fix a variable to a value it is removed even if it is specified in the protected list The routine returns a zero on success and a nonzero if an error occurs int CPXcopyprotected CPXCENVptr env CPXLPptr lp int cnt const int indices CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object
45. to existing routines in the Callable Library There are no advanced feature changes for Concert Technology ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 191 CPLEX 8 0 ADVANCED FEATURES RELEASE NOTES Conversion Notes The introduction of const arguments throughout the CPLEX Callable Library requires that user written callback routines be modified to use const arguments The function used to specify branching within a branch callback has been changed The function has been renamed from CPXbranchcallbackbranch to CPXbranchcallbackbranchbds to reflect that this is the particular branching function that creates a node by changing variable bounds A parameter has been added to allow the association of user data with the created node The function CPXget kappa has been changed to return an estimate of the condition number of simplex basis instead of the exact value The estimate can be computed more quickly than the exact condition number A new function CPXgetExactkappa returns the exact condition number as CPXgetkappa did in previous releases New features Exact Condition Number As noted in the Conversion Notes the existing Callable Library routine CPXgetkappa now returns an upper bound on the condition number of the simplex basis The exact condition number can be obtained by calling the new function CPXgetExactkappa Branching on Constraints Two new branching functions are provided for branching on c
46. whichinfo void result_p Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 93 CPXgetcallbacknodeinto 94 int wherefrom An integer value indicating where the user written callback was called from This parameter must be the value of wnerefrom passed to the user written callback int nodeindex The index of the node for which information is requested Nodes are indexed from 0 zero to nodecount 1 where nodecount is obtained from the callback information function CPXgetcallbackinfo with a whichinfo value of CPX_CALLBACK_INFO_NODES_LEFT int whichinfo An integer indicating which information is requested Table 1 summarizes possible values Table 1 Information Requested for a User Written Node Callback Symbolic Constant C Type Meaning CPX_CALLBACK_INFO_NODE_SIINF double sum of integer infeasibilities CPX_CALLBACK_INFO_NODE_NIINF int number of integer infeasibilities CPX_CALLBACK_INFO_NODE_ESTIMATE double estimated integer objective CPX_CALLBACK_INFO_NODE_DEPTH int depth of node in branch cut tree CPX_CALLBACK_INFO
47. 0 PXhybnetopt 31 PXkilldnorms 136 PXkillpnorms 137 PXmdleave 138 PXmipopt 22 46 48 50 67 70 151 154 162 163 PXpivot 140 PXpivotin 141 PXpivotout 143 PXpreaddrows 24 144 PXprechgobj 145 PXpresolve 22 23 146 PXprimopt 22 23 137 PXqpdjfrompi 147 PXqpuncrushpi 149 PXsetbranchcallbackfunc 29 150 PXsetcutcallbackfunc 28 154 ADVANCED REFERENCE MANUAL CPXsetdeletenodecallbackfunc 157 CPXsetheuristiccallbackfunc 27 159 CPXsetincumbentcallbackfunc 162 CPXsetnodecallbackfunc 31 165 CPXsetsolvecallbackfunc 31 167 CPXslackfromx 169 CPXsolution 65 66 92 100 134 149 179 180 CPXsolwrite 170 CPXstrongbranch 173 CPXtightenbds 175 CPXuncrushform 177 CPXuncrushpi 72 CPXuncrushx 149 169 180 CPXunscaleprob 181 C continued ctypes accessing from user written callback 84 cut adding 67 adding user 182 callbacks and 154 deleting user 183 dual reductions and 21 duplicates of 154 indexing 154 integer infeasibility and 154 naming 154 no deletion of 154 pool 154 right hand side RHS 67 70 sense of 68 71 sequence numbers 154 subproblem and 154 cut callback 28 cutting plane algorithms 141 D deleting lazy constraints 183 user cuts 183 detecting redundant row 41 INDEX diverging column 121 row 121 dual solution crushing original problem to presolved 65 uncrushing presolved problem to original 149 179 dual steepest edge iterations 173 dual steepest edge norms 58 79
48. 73 steepest edge norms 58 61 79 118 127 141 strengthening bounds 41 subproblem accessing pointer to 100 cut off 159 cuts and 116 154 modifying 100 optimal basis of 160 optimizing 135 167 primal solution values 160 U unbounded problem 121 unscaling 181 upper bound values accessing for subproblem from user written callback 104 accessing from user written callback 88 user cut adding 182 clearing 77 deleting 183 pool 33 to 36 UserCutCallbackI nested class of 110Cp1ex 187 V variable accessing priority of 115 accessing set that cannot be aggregated out 131 basic 55 175 branching 173 branching at node 94 152 diverging 121 fixed 143 173 175 in resident basis 81 pivoting out 143 primal steepest edge norms 127 protecting from substitution 62 206 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL INDEX WwW writing solutions 170 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 207 INDEX 208 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL
49. ACK 32 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL User Cut and Lazy Constraint Pools Sometimes a user may know a large set of cutting planes a priori user cuts or have additional constraints that are unlikely to be violated lazy constraints Simply including these cuts or constraints in the original formulation would make the LP subproblem of a MIP optimization very large and or expensive to solve Instead these can be handled through the cut callback described in Chapter 2 Advanced MIP Control Interface or by setting up cut pools before MIP optimization begins 00q JUIB1ISUOD Cc 7 ad e D 3 Q m N lt This chapter contains the sections Adding User Cuts and Lazy Constraints Deleting User Cuts and Lazy Constraints Adding User Cuts and Lazy Constraints You may add user cuts or lazy constraints through add routines in the Component Libraries or via LP and SAV files ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 33 ADDING USER CUTS AND LAZY CONSTRAINTS Component Libraries The following routines will add to the user cut pool 4 The CPLEX Callable Library routine is CPXaddusercuts CPLEX will scan the cut pool for violated cuts and add these to the LP subproblem and no integer solution will violate the cuts The Concert Technology routine is IloCplex addUserCuts The following routines will add to the lazy constraint pool 4 The CPLEX Callable Library routine is CPX
50. ACK6 name typel xl type2 x2 type3 x3 type4 x4 type5 x5 type6 x6 ILOLAZYCONSTRAINTCALLBACK7 name typel xl type2 x2 type3 x3 type4 x4 type5 x5 type6 x6 type 7 x7 This macro creates two things an implementation class for a user defined lazy constraint callback named namelI and a function named name that creates an instance of this class and returns an IloCplex Callback handle for it This function needs to be called with an environment as first parameter followed by the n parameters specified at the macro execution in order to create a callback The callback can then be used passing it the use method of an I10Cplex object The class name that is created by the macro includes the implementation of method makeClone as required for callbacks The implementation of method main must be provided by the user in parentheses following the macro invocation ILOLAZYCONSTRAINTCALLBACKn name implementation of the callback For the implementation of the callback methods from class IloCplex LazyConstraintCallbackI and its parent classes can be used You are not obliged to use this macro to define callbacks When the macro seems too restrictive for your purposes we recommend that you define a callback class directly Since the argument name is used to name the callback class it is not possible to use the same name for several callback definitions
51. ARAM RELAXPREIND 18 PX_PPRIIND_STEEP 137 PX_PRECOL_AGG 129 PX_PRECOL_FIX 129 PX_PRECOL_LOW 129 PX_PRECOL_OTHER 129 PX_PRECOL_UP 129 PX_PREREDUCE_DUALONLY 21 PX_P EDUCE_NO_PRIMALORDUAL 21 PX_P EDUCE_NOPRIMALORDUAL 39 T EJ Ty PX _P EDUCE PRIMA ANDDUAL 21 R R R R EDUCE PX P PX P E E a x E R e R R R R R RI RI ROW_AGG PRIMA 129 PX _P ROW_O THER 129 ED 129 ONLY 21 39 PX _PR ROW_R PX_SOS1 115 PX_SOS2 115 PX_TYPE_SOS1 94 152 PX_TYPE_SOS2 94 152 PX_TYPE_USER 94 152 PX_TYPE_VAR 94 152 PX_UNBOUNDED 103 133 PXaddcols 24 PXaddlazyconstraints 39 75 PXaddrows 39 40 144 PXaddusercuts 21 40 77 PXbaropt 22 PXbasicpresolve 24 41 PXbinvacol 42 PXbinvarow 43 PXbinvcol 44 PXbinvrow 45 PXbranchcallbackbranchbds 46 PXbranchcallbackbranchconstraints 48 50 PXbranchcallbackbranchgeneral 50 PXbt ran 53 201 INDEX CPXcheckax 54 CPXcheckpib 55 CPXchgbds 24 CPXchgob3 145 CPXchgrhs 24 CPXcloneprob 23 28 CPXcopybase 23 CPXcopybasednorms 56 79 CPXcopydnorms 58 CPXcopypartialbase 59 CPXcopypnorms 61 CPXcopyprotected 21 30 62 CPXcopystart 22 23 CPXcrushform 63 CPXcrushpi 65 CPXcrushx 66 CPXcutcallbackadd 21 28 67 154 CPXcutcallbackaddlocal 70 CPXdjfrompi 72 CPXdualfarkas 73 CPXd
52. ARAM_REDUCE is set to CPX_PREREDUCE_DUALONLY 2 presolve can conclude that problem is primal unbounded if it is primal feasible with return status CPXERR_PRESLV_UNBD ta A final facility that modifies the set of reductions performed by presolve is the CPXcopyprotected routine The user provides as input a list of variables in the original ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 21 MANUAL CONTROL OF PRESOLVE problem that should survive presolve that is should exist in the presolved problem Presolve will avoid reductions that remove those variables with one exception If a protected variable can be fixed presolve will still remove it from the problem This command is useful in cases where the user wants to explicitly control some aspect of the branch amp cut process for example through the branch callback using knowledge about those variables Manual Control of Presolve While presolve was a highly automated and transparent procedure in releases of CPLEX prior to 7 0 releases 7 0 and above give the user significant control over when presolve is performed and what is done with the result This section discusses these added control facilities Recall that the functions mentioned here are described in detail including arguments and return values in Part II Routines The first control function provided by the advanced presolve interface is CPXpre
53. B for both bounds 152 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsetbranchcallbackfunc int bd Together with indices and 1u this array defines the bound changes for each of the created nodes The entry bd i indicates the new value of the bound int useraction_p A pointer to an integer indicating the action for ILOG CPLEX to take at the completion of the user callback Table 6 summarizes the possible actions Table 6 Actions to be Taken After a User Written Branch Callback Value Symbolic Constant Action 0 CPX_CALLBACK_DEFAULT Use CPLEX selected branch 1 CPX_CALLBACK_FAIL Exit optimization 2 CPX_CALLBACK_SET Use user selected branch as defined by calls to CPXbranchcallbackbranchbds 3 CPX_CALLBACK_NO_SPACE Allocate more space and call callback again Example status CPXsetbranchcallbackfunc env mybranchfunc mydata See Also The example admipexl c CPXgetbranchcallbackfunc CPXbranchcallbackbranchbds Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 153 CPXsetcutcallbackfunc CPXsetcutcallbackfunc Usage Description Return Value Synopsis 154 Mixed Integer Users Only The routine CPXsetcutcallbackfunc sets and modifies the user written callback for adding cuts The user written callback is called by ILOG CPLEX during MIP branch amp cut for every node that has an LP optimal solution with objective value below the
54. CALLBACK_INFO_SOS_PRIORITY 115 CPX_CALLBACK_INFO_SOS_SIZE 115 CPX_CALLBACK_INFO_SOS_TYPE 115 CPX_CALLBACK_MIP 84 86 88 90 92 96 98 102 104 106 108 110 CPX_CALLBACK_MIP_BRANCH 84 86 88 90 92 96 98 100 102 104 106 108 110 114 151 163 CPX_CALLBACK_MIP_CUT 67 70 84 86 88 90 92 96 98 100 102 103 104 106 108 110 114 155 CPX_CALLBACK_MIP_DELETEMODE 158 CPX_CALLBACK_MIP_HEURISTIC 84 86 88 90 92 96 98 100 102 104 106 108 110 114 160 CPX_CALLBACK_MIP_INCUMBENT 84 86 88 90 92 96 98 102 104 106 108 110 CPX_CALLBACK_MIP_NODE 84 86 88 90 92 93 96 98 102 104 106 108 110 112 166 CPX_CALLBACK_MIP_SOLVE 84 86 88 90 92 98 100 104 108 110 168 CPX_CALLBACK_NO_SPACE 153 CPX_CALLBACK_SET 153 156 161 163 166 168 CPX_DPRIIND_FULLSTEEP 136 CPX_DPRIIND_STEEP 136 CPX_IMPLIED_INTEGER_FEASIBLE 97 CPX_INTEGER_FEASIBLE 96 CPX_INTEGER_INFEASIBLE 30 96 CPX_OPTIMAL 103 CPX_PARAM_BRDIR 109 CPX_PARAM_CUTSFACTOR 67 CPX_PARAM_DGRADIENT 58 ILOG CPLEX 8 1 Cr Fa OG 0601010100400 701000000 010000 OO A ADVANCED REFERENCE MANUAL PX_PARA PX_PARA 70 PX_PARA PX PARA INDEX DPRIIND 79 136 110 MI 84 15 PG PP PCBREDLP 27 28 29 30 46 48 50 67 86 88 90 92 96 98 104 106 108 1 154 162 163 RADIENT 61 RIIND 137 PX PARAM PRELINEAR 21 28 40 154 PX_PARAM REDUCE 21 22 24 29 39 76 134 144 145 154 PX P
55. CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_SOLVE CPX_CALLBACK_MIP_CUT or Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetcallbackorder CPXCENVptr env void cbdata int int int int int Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata wherefrom priority direction begin end The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback int priority An array where the priority values are to be returned This array must be of length at least end begin 1 If successful priority 0 through priority end begin contain the priority order values May be NULL 108 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbackorder int direction An array where the preferred direction values are to be returned This array must be of length at least end begin 1 If successful direction 0 through direction end begin contain the preferred direction values May be NULL The value of direction j will be CPX_BRANCH_GLOBAL 0 use global branching direction setting CPX
56. CED REFERENCE MANUAL 7 TABLE OF CONTENTS 8 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL About This Manual This manual documents advanced routines of the ILOG CPLEX Component Libraries version 8 1 It supplements the LOG CPLEX Reference Manual For Callable Library routines such as CPXgetcallbackinfo mentioned but not documented in this manual consult the ILOG CPLEX Reference Manual available on line in the standard distribution of the product The examples mentioned in this document are available by anonymous login at the CPLEX ftp site ftp ftp cplex com pub examples v81l advanced src in the subdirectory corresponding to the version number of the product The chapters in Part I Concepts provide background information on using the advanced presolve and MIP control routines as well as information on user cuts and lazy constraints and should be read before using those features All of the advanced routines are documented in alphabetical order in Part Il Routines of this manual They serve a variety of purposes as indicated below They should be considered subject to change in future releases in either functionality or calling sequence and possibly subject to removal if conditions warrant In addition advanced routines typically demand an understanding of the algorithms used by CPLEX and thus incur a heightened risk of incorrect behavior in your application program behavior that can be difficult to debug Therefore w
57. CPLEX Reference Manual int i An integer that indicates the index of the row to be computed double y An array containing the i row of B The array must be of length at least equal to the number of rows in the problem ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 45 CPXbranchcallbackbranchbds CPXbranchcallbackbranchbds Usage Mixed Integer Users Only Description The routine CPXbranchcallbackbranchbds specifies the branches to be taken from the current node It may be called only from within a user written branch callback function Branch variables are in terms of the original problem if the parameter CPX_PARAM MIPCBREDLP is set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise branch variables are in terms of the presolved problem Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXbranchcallbackbranchbds CPXCENVptr env void cbdata int wherefrom double nodeest int cnt int indices const char lu const int bd void userhandle int segnum_p Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPL routines ira x lt void cbdata A pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value that indicates where the user written callback was called from
58. CPXge tkappa ves ted bt ik edd a etd Deedee ome eee 124 CPXgetnodecallbackfunc 1 0 0 0 eee nee ents 125 CPXgetobjoffset 0 ete eee 126 CPXQeIPNOMMS ia sist a wah Dein ei ee eS Rdg eRe HE ad aw Sse Bane Se aden 127 CPXQO presta ic rnia costara A O Hie goes ee eee 128 CPXgeiprotected soa ci thas e a A AR ee ea ae eS 131 COPXgetlay iii ri dvi Gs e bap bine pa ead ponte a E 133 CPXgetredip inseri a id 134 CPXgetsolvecallbackfunc 0 cette eee 135 GCPXKINANONMS isie nace dead chap dcanneda para eens Sa ey Gk aes Ate eat EES aca 136 GPXKIIIPNONIMS eiu wis a8 Sar BA a it A ee argent as 137 CPXMAAVE is dec eee E eee eA eae Seta ee ena Ra ee 138 ICE XDIVOL o adanan iea a ci Ae necia Gane 140 CPXDIVOUN ri bad ce wh phd ie bin Wee weed pee he bees PANS TEESE Eee 141 GCPXDIVOTOUL 4 2 dienes Get ddd id dat tac Podge be 143 CPXpredddhOwS 2 snes ph ri waa ae gee Maken waka a 144 CPXprechgob ssi vgs hana o pee ed ees hee ee ee 145 A AAA PP ate este eigen deat A ae Beare a eet acetals Seen ee 146 CPXOPAOMP lings scn vee pees ee AR A Open eee Ree 147 CPXGPUNCKUSIPL vids fc hace Saas Hekate eee a ed Rea Sandan eae Pelee eRe eS 149 6 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL Appendix A TABLE OF CONTENTS CPXsetbranchcallbackfunc ooocoocoococooo tees 150 CPXsetcutcallbackfUnC 0 0 ete nent eee 154 CPXsetdeletenodecallbackfunC 0 cece tenets 157 CPXsetheuristiccallbackfUnC ooooocoocococco eee 159 CP
59. ENVptr env void cbdata int wherefrom double uppc double downpc int begin int end Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLI routines ie x lt void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback 110 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbackpseudocosts double uppc An array to receive the values of up pseudo costs This array must be of length at least end begin 1 If successful uppc 0 through uppc end begin will contain the up pseudo costs May be NULL double downpc An array to receive the values of the down pseudo costs This array must be of length at least end begin 1 If successful downpc 0 through downpc end begin will contain the down pseudo costs May be NULL int begin An integer indicating the beginning of the range of pseudo costs to be returned int end An integer indicating the end of the range of pseudo costs to be returned Example status CPXgetcallbackpseudocosts env cbhdata wherefrom upcost downcost j k ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 111 CPXgetcallbackseqinfo CPXgetcallbackseqinfo Us
60. EX 8 1 ADVANCED REFERENCE MANUAL 29 INCUMBENT CALLBACK information on these variables call CPXgetcallbackorder to get priority order information make a decision based on this and perhaps other information and then respond that the current node will have two children where one has a new lower bound on the branch variable and the other has a new upper bound on that variable Alternatively the branch callback routine can be used to sculpt the search tree by pruning nodes or adjusting variable bounds Choosing zero children for a node prunes that node while choosing one node with a set of new variable bounds adjusts bounds on those variables for the entire subtree rooted at this node Note that the user must be careful when using this routine for anything other than choosing a different variable to branch on Pruning a valid node or placing an invalid bound on a variable can prune the optimal solution We should point out one important detail associated with the use of the CPX_PARAM_MIPCBREDLP parameter in a branch callback If this parameter is set to CPX_OFF 0 the user can choose branch variables and add bounds for the original problem However not every fractional variable is actually available for branching Recall that some variables are replaced by linear combinations of other variables in the presolved problem Since branching involves adding new bounds to specific variables in the presolved problem a variable must be present i
61. G CPLEX 8 1 ADVANCED REFERENCE MANUAL 129 CPXgetprestat Example status CPXgetprestat env lp amp presolvestat precstat prerstat origcstat origrstat See Also Example admipex6 c in the advanced examples directory 130 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetprotected CPXgetprotected Usage Description Return Value Synopsis Arguments Advanced The routine CPXgetprotected is used to get the set of variables that cannot be aggregated out The routine returns a zero on success and a nonzero if an error occurs The value CPXERR_NEGATIVE_SURPLUS indicates that insufficient space was available in the array indices to hold the protected variable indices int CPXgetprotected CPXCENVptr env CPXCLPptr lp int cnt_p int indices int pspace int surplus_p CPXCENVptr env irl pS The pointer to the CPLEX environment as returned by one of the CPXopenCPLI routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int cnt_p A pointer to an integer to contain the number of protected variables returned that is the true length of the array indices int indices The array to contain the indices of the protected variables int pspace An integer indicating the length of the array indices int surplus_p A pointer to an integer to contain the difference between p
62. LOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetdeletenodecallbackfunc CPXgetdeletenodecallbackfunc Usage Description Return Value Synopsis Arguments Example See Also CPXC int void Mixed Integer Users Only The routine CPXgetdeletenodecallbackfunc accesses the user written callback routine to be called during MIP optimization when a node is to be deleted Nodes are deleted when a branch is carried out from that node when the node relaxation is infeasible or when the node relaxation objective value is worse than the cutoff This callback can be used to delete user data associated with a node For documentation of callback arguments see the routine CPXsetdeletenodecallbackfunc This routine does not return a result void CPXgetdeletenodecallbackfunc CPXCENVptr env int CPXPUBLIC deletenodecallback_p CPXCENVptr env void cbdata int wherefronm void cbhandle int segnum void handle void cbhandle_p ENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines E x lt CPXPUBLIC deletenodecallback_p The address of the pointer to the current user written delete node callback If no callback has been set the pointer evaluates to NULL cbhandle_p The address of a variable to hold the user s private pointer CPXgetdeletenodecallbackfunc env amp current_callback amp current_handle CPXsetdeletenodec
63. P Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 115 CPXgetcutcallbackfunc CPXgetcutcallbackfunc Usage Description Return Value Synopsis Arguments Example See Also 116 Mixed Integer Users Only The routine CPXgetcutcallbackfunc accesses the user written callback for adding cuts The user written callback is called by ILOG CPLEX during MIP branch amp cut for every node that has an LP optimal solution with objective value below the cutoff and that is integer infeasible The callback routine adds globally valid cuts to the LP subproblem For documentation of callback arguments see the routine CPXsetcutcallbackfunc This routine does not return a result void CPXgetcutcallbackfunc CPXCENVptr env int callback_p CPXCENVptr xenv void cbdata int wherefrom void cbhandle int useraction_p void cbhandle_p CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines sal x lt int callback_p The address of the pointer to the current user written cut callback If no callback has been set the pointer evaluates to NULL void cbhandle_p The address of a variable to hold the user s private pointer CPXgetcutcallbackfunc env amp current_cutfunc amp current_data CPXcutcallbackadd CPXgetcallbacksosinfo CPXgetcallbacknodelp CPXsetcutcallbackfunc Advanced MIP Control Interface I
64. PLI routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int plen The number of entries in the arrays pind and pval const int pind const double pval The linear formula in terms of the presolved problem Each entry pind i indicates the column index of the corresponding coefficient pval i int len_p A pointer to an integer to receive the number of nonzero coefficients that is the true length of the arrays ind and val double offset_p A pointer to a double to contain the value of the linear formula corresponding to variables that have been removed in the presolved problem ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 177 CPXuncrushform int ind double val Example The linear formula in terms of the original problem Let cols CPXgetnumcols env 1p Ifind i lt cols then the 1 variable in the formula is variable with index ind i in the original problem If ind i gt cols then the it variable in the formula is the slack for the ina 1 cols ranged row The arrays ind and val must be of length at least the number of columns plus the number of ranged rows in the original LP problem object status CPXuncrushform env lp plen pind pval amp len amp offset ind val 178 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXuncrushpi CPXuncrushpi Usage Description Return Value Synopsi
65. PXgetnumcols int rstat An array containing the basis status of the rows in the constraint matrix The length of the allocated array is at least the value returned by CPXgetnumrows double dnorm An array containing the dual steepest edge norms The length of the allocated array is at least the value returned by CPXgetnumrows CPXcopybasednorms CPXgetnumcols CPXgetnumrows ILOG CPLEX Reference Manual ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetbhead CPXgetbhead Usage Description Return Value Synopsis Arguments CPXCI Advanced The routine CPXgetbhead returns the basis header it gives the negative value minus one of all row indices of slacks The routine returns a zero on success and a nonzero if an error occurs int CPXgetbhead CPXCENVptr env CPXCLPptr lp int head double x ENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPL routines E x lt CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int head An array containing the indices of the variables in the resident basis where basic slacks are specified by the negative of the corresponding row index minus 1 one that is rowindex 1 The array must be of length at least equal to the number of rows in the LP problem object double x ILOG An array containing the values
66. TING ROUND OFF ERRORS tcallbac tcallbac tcallbac tcallbac tcallbac Llbac Llbac llbac llbac Llbac llbac lLlbac Llbac kincumbent k1p knodeinfo knodeintfeas knodelb knodelp knodeobjval knodeub knodex korder kpseudocosts kseqinfo ksosinfo Detecting Round Off Errors in a Solution Using the Basis cPXcheckax CPxcheckpib 9 SE SEE SEE JEE SEE 445 4 4 00000000000 PXget PXget PXbtran PXftran PXbinvacol PXbinvarow PXbinvcol PXbinvrow PXgetbhead PXgetijrow kappa Exactkappa IN A SOLUTION ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 11 ACCESSING AND MANIPULATING NORMS Accessing and Manipulating Norms PXgetbasednorms PXcopybasednorms PXgetdnorms PXcopydnorms PXkilldnorms PXgetpnorms PXcopypnorms 9 96 2 900 000 e E e e e e E PXkillpnorms Accessing Information About the Current Problem CPXgetijdiv cPXgetobjoffset Changing the Basis by Pivoting CPXpivot CPXpivotin CPXpivotout Using User Cuts and Lazy Constraints CPXaddusercuts CPXfreeusercuts CPXaddlazyconstraints CPXfreelazyconstraints Advanced Presolve Routines CPXbasicpresolve 12 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 9 96 189 909 999 9 49 4 49 49 4 MISCELLANEOUS ROUTINES PXcopyprotected PXcrushform PXcrushpi PXcrushx PXfreepresolve PXgetprestat PXgetprotected PXgetredlp PXpreaddrows PXprechgobj PXpreso
67. X_CALLBACK_MIP_HEURISTIC Or CPX_CALLBACK_MIP_CUT The routine returns a zero on success and a nonzero if an error occurs int CPXgetcallbacknodeobjval CPXCENVptr env void cbdata int wherefrom double objval_p CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback double objval_p A pointer to a variable of type double where the objective value of the node subproblem 1s to be stored status CPXgetcallbacknodeobjval env cbdata wherefrom amp objval Examples admipex1 c and admipex3 c in the advanced examples directory ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbacknodestat CPXgetcallbacknodestat Usage Description Return Value Synopsis Arguments Example Mixed Integer Users Only The routine CPXgetcallbacknodestat 1s used to get the optimization status of the subproblem at the current node from within a user written callback during MIP optimization The optimization status will be either optimal or unbounded An unbounded status can occur when some of the constraints are being t
68. Xgetbasednorms CPXcopybasednorms copies the values in the arrays cstat rstat and dnorm as returned by CPXgetbasednorms into a specified problem object Each of the arrays cstat rstat and dnorm must be non NULL Only data returned by CPXgetbasednorms should be copied by CPXcopybasednorms Other details of estat rstat and dnorm are not documented Important The routine CPXcopybasednorms should be called only if the return values of CPXgetnumrows and CPXgetnumcols have not changed since the companion call to CPXgetbasednorms If either of these values has increased since that companion call a memory violation may occur If one of those values has decreased the call will be safe but its meaning will be undefined The routine returns a zero on success and a nonzero if an error occurs double CPXcopybasednorms CPXCENVptr env CPXLPptr lp const int cstat const int rstat const double dnorm CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines E x lt CPXLPptr lp A pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual const int cstat An array containing the basis status of the columns in the constraint matrix returned by a call to CPXgetbasednorms The length of the allocated array must be at least the value returned by CPXgetnumcols const int rstat An
69. XsetincumbentcallbackfUNC oooococccccocor eee 162 CPXsetnodecallbackfUnC oooocococococc een eee 165 CPXsetsolvecallbackfUNC 6 ete eet eee 167 CPXslackifomXx 2 cois eed De ies dies Debden Padi oh ded ie Pe ae ae eee 169 CPXsOlWrite s a c ch seen biased A eh eee eth Re A mr seks 170 CPXstrongbranchs nesai sarah ee RRA Reed a ase Pe ae eee ee Oe Pe 173 CPXtightenbds 2 20 cc wi piadiwed bia eee OE eae pone baad eee 175 CPXUNCHISHIONNT cios bb esa cae eda caked hese be he ed ad betwee 177 CPXUNCMSNPL sia ade RE EMSA Babee ioe ke 179 CPXUNCHUSHX 5 acess A ai aca 180 CPXUNSCaleprob viii Fe ete otis ee Sh Rare A Sed Mae ee Nadine 181 NOCDIEX ates cette cee Meee eee ee DR aR EE 182 lloCplex LazyConstraintCallbackl 2 0 0 186 lloCplex UserCutCallbackl 0 0 ete tenes 187 ILOLAZYCONSTRAINTCALLBACK 0 000 c cece eee eee eee eee 188 ILOUSERCUTCALLBACK 2 22 2 00000 ccc c hi a a ERR EARE E eG 189 Advanced Features Release NoteS 2 00 cece eee eee eee 191 CPLEX 8 1 Advanced Features Release NoteS 0 0 cece cece eee os 191 CPLEX 8 0 Advanced Features Release NoteS 00 000 cee ee eee eee eee 191 CPLEX 7 1 Advanced Features Release NoteS 0 0 ce cece eee e eee eee 193 CPLEX 6 6 to 7 0 Advanced Features Release Notes 00 0c eee ee eee coc 196 A A A kee Gece ee et da Siac Shee Aare eae eae eee ates eth 197 E E esc eG io a as 199 ILOG CPLEX 8 1 ADVAN
70. _NODE_OBJVAL double objective value of LP subproblem CPX_CALLBACK_INFO_NODE_TYPE int type of branch at this node see Table 2 CPX_CALLBACK_INFO_NODE_VAR int for nodes of type CPX_TYPE_VAR the branching variable for this node for SOS type branches the rightmost variable in left subset CPX_CALLBACK_INFO_NODE_SOS int the number of the SOS used in branching 1 if none used CPX_CALLBACK_INFO_NODE_SEQNUM int sequence number of the node CPX_CALLBACK_INFO_NODE_NODENUM int node index of the node Table 2 summarizes possible values returned when the type of information requested is branch type that is wnichinfo CPX_CALL Table 2 Branch Types Returned when whichinfo BACK_INFO_NODE_TYPE CPX_CALLBACK_INFO_NODE_TYPE Symbolic Constant Value Branch Type CPX_TYPE_VAR ugi variable branch CPX_TYPE_SOS1 ra SOS1 branch CPX_TYPE_SOS2 r21 SOS2 branch CPX_TYPE_USER xX user defined ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbacknodeinto void result_p A generic pointer to a variable of type double or int representing the value returned by whichinfo The column C Type in Table 1 indicates the type of various values returned by whichinfo Example status CPXgetnodecallbackinfo env curlp wherefrom 0 CPX_CALLBACK_INFO_NODE_NIINF amp numiinf See Also CPXgetcallbackinfo ILOG CPLEX Reference Manual CPXgetcallbackseqinfo
71. _PARAM_BRDIR CPX_BRANCH_DOWN 1 branch down first on variable j begin CPX_BRANCH_UP 1 branch up first on variable j begin int begin An integer indicating the beginning of the range of priority order information to be returned int end An integer indicating the end of the range of priority order information to be returned Example status CPXgetcallbackorder env cbdata wherefrom priority NULL 0 cols 1 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 109 CPXgetcallbackpseudocosts CPXgetcallbackpseudocosts Usage Mixed Integer Users Only Description The routine CPXgetcallbackpseudocosts is used to get the pseudo cost values during MIP optimization from within a user written callback The values are from the original problem if CPX_PARAM_MIPCBREDLP is set to CPX_OFF Otherwise they are from the presolved problem Note When pseudo costs are retrieved for the original problem variables pseudo costs are zero for variables that have been removed from the problem since they are never used for branching This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_SOLVE or CPX_CALLBACK_MIP_CUT Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetcallbackpseudocosts CPXC
72. a lazy constraint in the LP as soon as it becomes violated In other words the solution computed by IloCplex ensures that all the lazy constraints that have been added are satisfied By contrast if the constraint does not change the feasible region of the extracted model but only strengthens the formulation it is referred to as a user cut While user cuts can be added to IloCplex with addLazyConstraint it is generally preferable to do so using addUserCuts It is an error however to add lazy constraints using addUserCuts When columns are deleted from the extracted model all lazy constraints are deleted as well and need to be recopied into the lazy constraint pool Use of this method in place of addCuts allows for further presolve reductions IloRangeArray addLazyConstraints IloRangeArray rng This member function adds a set of lazy constraints to the invoking I1oCplex object Everything said for addLazyConstraint applies to each of the lazy constraints given in array rng IloRange addUserCut IloRange rng This member function adds rng as a user cut to the invoking IloCplex object The range rng is copied into the user cut pool the rng itself is not part of the pool so changes to rng after it has been copied into the user cut pool will not affect the user cut pool 182 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL Cuts added with addUserCut must be real cuts in that the solution of a MIP does not depend on whether the cuts ar
73. ackglobalub int end An integer indicating the end of the range of upper bound values to be returned Example status CPXgetcallbackglobalub env cbdata wherefrom gub 0 cols 1 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 89 CPXgetcallbackincumbent CPXgetcallbackincumbent Usage Mixed Integer Users Only Description The routine CPXgetcallbackincumbent is used to get the incumbent values during MIP optimization from within a user written callback The values are from the original problem if CPX_PARAM_MIPCBREDLP is set to CPX_OFF otherwise they are from the presolved problem This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_SOLVE or CPX_CALLBACK_MIP_CUT Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetcallbackincumbent CPXCENVptr env void cbdata int wherefrom double x int begin int end Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was calle
74. addlazyconstraints CPLEX will scan the pool for violated constraints and add these to the MIP subproblems and no integer solution will violate the constraints Additionally the tight lazy constraints will be added to the fixed problem obtained by calling the routine CPXchgprobtype 4 The Concert Technology routine is IloCplex addLazyConstraints Note that CPLEX does not guarantee that user cuts and lazy constraints are added as soon as they are violated by a node relaxation It simply guarantees that no feasible solutions returned by CPLEX will violate these constraints Reading LP and SAV Files User cuts and lazy constraints may also be specified in LP format and SAV format files and so may be read With the Interactive Optimizer Through the routines CPXreadcopyprob and IloCplex importModel General Syntax The general syntax rules for LP format given in the ILOG CPLEX Reference Manual apply to user cuts and lazy constraints 4 The user cuts section or sections must be preceded by the keywords USER CUTS The lazy constraints section or sections must be preceded by the keywords LAZY CONSTRAINTS These sections and the ordinary constraints section preceded by the keywords SUBJECT TO can appear in any order and can be present multiple times as long as they are placed after the objective function section and before any of the keywords BOUNDS GENERALS BINARIES SEMI CONTINUOUS or END
75. age Mixed Integer Users Only Description The routine CPXgetcallbackseqinfo accesses information about nodes during the MIP optimization from within user written callbacks This routine may be called only when the value of its wherefrom argument is CPX_CALLBACK_MIP_NODE The information accessed from this routine can also be accessed with the routine CPXgetcallbacknodeinfo Nodes are not stored by sequence number but by node number so using the routine CPXgetcallbackseqinfo can be much more time consuming than using the routine CPXgetcallbacknodeinfo A typical use of this routine would be to obtain the node number of a node for which the sequence number is known and then use that node number to select the node with the node callback Note This routine cannot retrieve information about nodes that have been moved to node files For more information about node files see the CPLEX User s Manual If the argument seqnun refers to a node in a node file CPXgetcallbacknodeinfo returns the value CPXERR_NODE_ON_DISK Return Value The routine returns a zero on success and a nonzero if an error occurs The return value CPXERR_NODE_ON_DISK indicates an attempt to access a node currently located in a node file on disk Synopsis int CPXgetcallbackseqinfo CPXCENVptr env void cbdata int wherefrom int segnum int whichinfo void result_p Arguments CPXCENVptr env The pointer to the ILOG CPLEX env
76. al norms available before the optimizer can begin If these norms are not already resident they must be computed ant that computation may be expensive The functions CPXgetbasednorms and CPXcopybasednorms can in some cases avoid that expense Suppose for example that in some application an LP is solved by CPXdualopt with one of those pricing settings After the solution of the LP some intermediate optimizations are carried out on the same LP and those subsequent optimizations are in turn followed by some changes to the LP and a re solve In such a case copying the basis and norms that were resident before the intermediate solves back into ILOG CPLEX data structures can greatly increase the speed of the re solve The routine returns a zero on success and a nonzero if an error occurs ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 79 CPXgetbasednorms Synopsis Arguments See Also 80 int CPXgetbasednorms CPXCENVptr env CPXCLPptr lp int cstat int rstat double dnorm CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines E x lt CPXCLPptr lp A pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int cstat An atray containing the basis status of the columns in the constraint matrix The length of the allocated array is at least the value returned by C
77. allback is set to NULL no callback is called during optimization void cbhandle Callback Description Return Value A pointer to user private data This pointer is passed to the callback ILOG CPLEX calls the solve callback before ILOG CPLEX solves the subproblem defined by the current node The user can choose to solve the subproblem in the solve callback instead by setting the user action parameter of the callback The optimization that the user provides to solve the subproblem must provide a ILOG CPLEX solution That is the ILOG CPLEX Callable Library routine cPxget stat documented in the ILOG CPLEX Reference Manual must return a nonzero value The user may access the Ip pointer of the subproblem with the Callable Library routine CPXgetcallbacknodelp The callback returns a zero on success and a nonzero if an error occurs ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 167 CPXsetsolvecallbackfunc Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata A pointer passed from the optimization routine to the user written callback that identifies the problem being optimized The only purpose of this pointer is to pass it to the callback information routines int wherefrom An integer value indicating where in the optimization this function was called It will have the value CPX_CALLBACK_MIP_SOLVE void cbhandle A pointer to
78. allbackfunc CPXbranchcallbackbranchbds CPXbranchcallbackbranchconstraints CPXbranchcallbackbranchgeneral Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 117 CPXgetdnorms Usage Description Return Value Synopsis Arguments See Also 118 Advanced The routine CPXgetdnorms accesses the norms from the dual steepest edge As in CPXcopydnorms the argument head is an array of column or row indices corresponding to the array of norms Column indices are indexed with nonnegative values Row indices are indexed with negative values offset by 1 one For example if head 0 5 norm 0 is associated with row 4 The routine returns a zero on success and a nonzero if an error occurs int CPXgetdnorms CPXCENVptr env CPXCLPptr lp double norm int head int len_p CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double norm An array containing the dual steepest edge norms in the ordered specified by head The array must be of length at least equal to the number of rows in the LP problem object int head An array containing column or row indices The allocated length of the array must be at least equal to the number of rows in the LP problem object int len_p
79. alues of cut coefficients Example status CPXcutcallbackaddlocal env cbdata wherefrom mynzcnt myrhs L P mycutind mycutval See Also CPXcutcallbackadd CPXgetcutcallbackfunc CPXsetcutcallbackfunc ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 71 CPXdjfrompi CPXdjfrompi Usage Advanced Description The routine CPXdjfrompi computes an array of reduced costs from an array of dual values This routine is for linear programs Use CPXqpdjfrompi for quadratic programs Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXdjfrompi CPXCENVptr env CPXCLPptr lp const double pi double dj Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCP LEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual const double pi An array that contains dual solution pi values for the problem as returned by routines such as CPXuncrushpi and CPXcrushpi The array must be of length at least the number of rows in the problem object double dj An array to receive the reduced cost values computed from the pi values for the problem object The array must be of length at least the number of columns in the problem object Example status CPXdjfrompi env lp pi dj 72 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL
80. amples directory 66 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXcutcallbackadd CPXcutcallbackadd Usage Mixed Integer Users Only Description The routine CPXcutcallbackadd adds cuts during MIP branch amp cut This routine may be called only from within user written cut callbacks thus it may be called only when the value of its wnerefrom argument is CPX_CALLBACK_MIP_CUT The cut may be for the original problem if the parameter CPX_PARAM_MIPCBREDLP was set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise the cut is used on the presolved problem Return Value The routine returns a zero on success and a nonzero if an error occurs One possible error is indicated by the symbolic constant CPXERR_NO_SPACE That error occurs when the number of cuts added reaches the maximum allowed as set by the parameter CPX_PARAM_CUTSFACTOR Synopsis int CPXcutcallbackadd CPXCENVptr env void cbdata int wherefrom int nzent double rhs int sense const int cutind const double cutval Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value that indicates where the user written callback was called from This parameter must be the value of wne
81. anch Types Returned when whichinfo CPX_CALLBACK_INFO_NODE_ TYPE 94 Table 3 Information Requested for a User Written SOS Callback 115 Table 4 SOS Types Returned when whichinfo CPX_CALLBACK_INFO_SOS TYPE 115 Table 5 Branch Types Returned from a User Written Branch Callback 152 Table 6 Actions to be Taken After a User Written Branch Callback 153 Table 7 Actions to be Taken After a User Written Cut Callback 155 Table 8 Actions to be Taken after a User Written Heuristic Callback 161 Table 9 Actions to be Taken after a User Written Incumbent Callback 163 Table 10 Actions to be Taken after a User Written Node Callback 166 Table 11 Actions to be Taken after a User Written Solve Callback 168 Table 12 Arguments of the write_entry Function in the Routine CPXsolwrite 171 Table 13 Bound Indicators in the Argument lu of CPXtightenbds 176 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 197 LIST OF TABLES 198 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL A accessing basis header 81 branch callback information 82 condition number estimate 124 ctypes from user written callback 84 divergent column 121 divergent constraint 121 divergent row 121 divergent variable 121 dual steepest edge norms 79 118 incumbent values from user written callback 90 informatio
82. ariable is not specified it is made basic This routine returns a zero on success and a nonzero if an error occurs int CPXcopypartialbase CPXCENVptr env CPXLPptr lp int cent const int cindices const int cstat int rent const int rindices const int rstat CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int cent An integer that indicates the number of variable column statuses specified and is the length of the cindices and cstat arrays const int cindices An array of length ccnt that contains the indices of the variables for which statuses are being specified const int cstat An array of length ccnt where the i entry contains the status for variable cindices i Values for cstat i CPX_AT_LOWER o variable at lower bound CPX_BASIC 1 variable is basic ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 59 CPXcopypartialbase CPX_AT_UPPER 2 variable at upper bound CPX_FREE_SUPER 3 variable free and non basic int rent An integer that indicates the number of slack surplus artificial statuses specified and is the length of the rindices and rstat arrays const int rindices An array of length rent that contains the indices of the slack surplus artificials for which statuses are bein
83. array containing the basis status of the rows in the constraint matrix returned by a call to CPXgetbasednorms The length of the allocated array must be at least the value returned by CPXgetnumrows ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXcopybasednorms const double dnorm An array containing the dual steepest edge norms returned by a call to CPXgetbasednorms The length of the allocated array must be at least the value returned by CPXgetnumrows See Also CPXgetbasednorms CPXgetnumcols CPXgetnumrows ILOG CPLEX Reference Manual ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 57 CPXcopydnorms CPXcopydnorms Usage Description Return Value Synopsis Arguments See Also 58 Advanced The routine CPXcopydnorms copies the dual steepest edge norms to the specified LP problem object The argument head is an array of column or row indices corresponding to the array of norms Column indices are indexed with nonnegative values Row indices are indexed with negative values offset by 1 one For example if head 0 5 then norm 0 is associated with row 4 The routine returns a zero on success and a nonzero if an error occurs int CPXcopydnorms CPXC CPXL cons cons int CPXCENVptr env ENVptr env Pptr lp t double norm t int head len The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A poi
84. as returned by CPXcreateprob documented in the CPLEX Reference Manual int cnt The number of variables to be protected const int indices An array of length cnt containing the column indices of variables to be protected from being substituted out status CPXcopyprotected env lp cnt indices ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXcrushtorm CPXcrushform Usage Advanced Description The routine CPXcrushform crushes a linear formula of the original problem to a Return Value Synopsis Arguments linear formula of the presolved problem The routine returns a zero on success and a nonzero if an error occurs int CPXcrushform CPXCENVptr env CPXCLPptr lp int len const int ind const double val int plen_p double poffset_p int pind double pval CPXCENVptr env irl pS The pointer to the CPLEX environment as returned by one of the CPXopenCPLI routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int len The number of entries in the arrays ind and val const int ind const double val The linear formula in terms of the original problem Each entry ind i indicates the column index of the corresponding coefficient val i int plen_p A pointer to an integer to receive the number of nonzero coefficients that is the true length of the arrays pind and pval double poffset_p
85. backgloballb and CPXgetcallbackglobalub for example return the tightest known global lower and upper bounds on all the variables in the problem No feasible solution whose objective is better than that of the best known solution can violate these bounds Similarly the routines CPXgetcallbacknodelb and CPXgetcallbacknodeub return variable bounds at this node These reflect the bound adjustments made during branching The routine CPXgetcallbackincumbent returns the current incumbent the best known feasible solution The routine CPXgetcallbacklp returns a pointer to the MIP problem presolved or unpresolved depending on the CPX_PARAM_MIPCBREDLP parameter This pointer can be used to obtain various information about the problem variable types etc or as an argument for the advanced presolve interface if the user wishes to manually translate between presolved and unpresolved values In addition the callback can use the codata parameter passed to it along with routine CPXgetcallbacknodelp to obtain a pointer to the node relaxation LP This can be used to access desired information about the relaxation row count column count etc Note that in both cases the user should never use the pointers obtained from these callbacks to modify the associated problems ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 27 CUT IA BANG K Cut Callback 28 As noted earlier the CPX_PARAM_MIPCBREDLP parameter influences the arguments t
86. be associated with the node created by this branch May be NULL int seqnum_p A pointer to an integer that on return will contain the sequence number that CPLEX has assigned to the node created from this branch The sequence number may be used to select this node in later calls to the node callback ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 49 CPXbranchcallbackbranchgeneral CPXbranchcallbackbranchgeneral Description Return Value Synopsis Arguments 50 The routine CPXbranchcallbackbranchgeneral specifies the branches to be taken from the current node when the branch includes variable bound changes and additional constraints It may be called only from within a user written branch callback function Branch variables are in terms of the original problem if the parameter EDLP is set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise branch variables are in terms of the presolved problem CPX_PARAM_MIPCBR The routine returns a zero on success and a nonzero if an error occurs int CPXbranchcallbackbranchgeneral CPXC void cbdata int wherefrom double nodeest int varcnt CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopencPLI routines void cbdata cons cons cons int int cons cons cons cons cons in Chee ct in rent nzcn ch in te Gk ch ere et int double rmatval ENVptr env t varin
87. bject Each of the arrays cstat rstat and dnorm must be non NULL That is each of these arrays must be allocated The allocated size of cstat is assumed by this routine to be at least the number returned by CPXgetnumcols The allocated size of rstat and dnorm are assumed to be at least the number returned by CPXgetnumrows Other details of cstat rstat and dnorm are not documented Success Failure If this routine succeeds cstat and rstat contain information about the resident basis and dnorm contains the dual steepest edge norms If there is no basis or if there is no set of dual steepest edge norms this routine returns an error code The returned data are intended solely for use by CPXcopybasednorms Example For example if a given LP has just been successfully solved by the ILOG CPLEX Callable Library optimizer CPXdualopt with the dual pricing option CPX_PARAM_DPRIIND set to CPX_DPRIIND_STEEP CPX_DPRIIND_FULLSTEEP or CPX_DPRIIND_STEEPQSTART then a call to CPXgetbasednorms should succeed That optimizer and those pricing options are documented in the LOG CPLEX Reference Manual and their use is illustrated in the ILOG CPLEX User s Manual Motivation When the ILOG CPLEX Callable Library optimizer CPXdualopt is called to solve a problem with the dual pricing option CPX_PARAM_DPRIIND set to CPX _DPRIIND_STEEP or CPX_DPRIIND_FULLSTEEP there must be values of appropriate du
88. callback to be called by ILOG CPLEX during MIP optimization after the subproblem has been solved to optimality That callback is not called when the subproblem is infeasible or cut off The callback supplies ILOG CPLEX with heuristically derived integer solutions If a linear program must be solved as part of a heuristic callback make a copy of the node LP and solve the copy not the CPLEX node LP The routine returns a zero on success and a nonzero if an error occurs int CPXsetheuristiccallbackfunc CPXENVptr env int CPXPUBLIC heuristiccallback CPXCENVptr env void cbdata int wherefrom void cbhandle double objval_p double x int checkfeas_p int useraction_p void cbhandle CPXENVptr env int void The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXPUBLIC heuristiccallback A pointer to a user written heuristic callback If this callback is set to NULL no callback is called during optimization cbhandle A pointer to user s private data This pointer is passed to the callback The call to the heuristic callback occurs after an optimal solution to the subproblem has been obtained The user can provide that solution to start a heuristic for finding an integer solution The integer solution provided to ILOG CPLEX replaces the incumbent ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 159 CPXsetheuristiccallbackfunc Return Value Argum
89. cumbent Callback A new callback the incumbent callback has been added so that users may examine integer feasible solutions to decide if they should replace the current incumbent This callback allows discarding solutions that do not meet criteria additional to that expressed in the mixed integer program formulation The incumbent callback together with the additional calls to the branch callback for integer feasible nodes provide a means to specify logical constraints instead of introducing indicator variables For example the user could accept only solutions where at least two variables from a set of continuous variables must be nonzero Solutions not meeting this criteria could be rejected in an incumbent callback and additional branches changing the bounds on these variables to nonzero values could be specified by a branch callback See the new routines CPXsetincumbentcallbackfunc CPXgetincumbentcallbackfunc User Cuts In the Concert Technology Library the following are added to provide capabilities for user cuts IloCplex addUserCut IloCplex addUserCuts IloCplex clearUserCuts IloCplex UserCutCallbackI ILOUSERCUTCALLBACK macro Lazy Constraints Two new functions have been added for the Callable Library CPXaddlazyconstraints cPXfreelazyconstraints They are used for constraints that are not implied by the constraints in the MIP problem In the prior version 7 0 these constraints co
90. d char varlu t varbd r double rhs ar sense t rmatbeg rmatind void userhandle int seqnum_p E gt A pointer passed to the user written callback This parameter must be the value of cbdata passed to the int wherefrom user written callback An integer value that indicates where the user written callback was called from This parameter must be the value of wherefrom passed to the user written callback double nodeest A double that indicates the value of the node estimate for the node to be created with this branch The node estimate is used to select nodes from the branch amp cut tree with certain values of the NodeSel parameter ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXbranchcallbackbranchgeneral int varcnt An integer that indicates the number of bound changes that are specified in the arrays varind varlu and varbd const int varind Together with varlu and varba this array defines the bound changes for the branch The entry varind i is the index for the variable const char varlu Together with varind and varba this array defines the bound changes for the branch The entry varlu i is one of three possible values indicating which bound to change L for lower bound vu for upper bound or 8 for both bounds const int varbd Together with varind and varlu this array defines the bound changes for the branch The entry varbd i indicates
91. d The parameter must be the value of wherefrom passed to the user written callback double x An atray to receive the values of the incumbent best available integer solution This array must be of length at least end begin 1 If successful x 0 through x end begin contain the incumbent values int begin An integer indicating the beginning of the range of incumbent values to be returned 90 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbackincumbent int end An integer indicating the end of the range of incumbent values to be returned Example status CPXgetcallbackincumbent env cbhdata wherefrom bestx 0 cols 1 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 91 CPXgetcallbacklp CPXgetcallbacklp Usage Mixed Integer Users Only Description The routine CPXgetcallback1p is used to get the pointer to the MIP problem that is in use when the user written callback function is called It is the original MIP if CPX_PARAM_MIPCBREDLP is set to CPX_OFF otherwise it is the presolved MIP In contrast the function CPXgetcallbacknodelp returns a pointer to the node subproblem which is an LP Generally this pointer may be used only in CPLEX Callable Library query routines such as CPXsolution or CPXgetrows This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE
92. d its associated presolved LP problem object Note that the CPLEX parameter CPX_PARAM_REDUCE must be set to CPX_PREREDUCE_PRIMALONLY 1 or CPX_PREREDUCE_NOPRIMALORDUAL 0 at the time of the presolve in order to change objective coefficients and preserve the presolved problem This routine should be used in place of CPXchgobj when it is desired to preserve the presolved problem Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXprechgobj CPXENVptr env CPXLPptr lp int cnt int indices double values Arguments The arguments and operation of CPXprechgobj are the same as those of CPXchgobj The objective coefficient changes are applied to both the original LP problem object and the associated presolved LP problem object Example status CPXprechgobj env lp objcnt objind objval See Also Example adpreex1 c in the advanced examples directory ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 145 CPXpresolve CPXpresolve Usage Description Return Value Synopsis Advanced The routine CPXpresolve performs LP or MIP presolve depending whether a problem object is an LP or a MIP If the problem is already presolved the existing presolved problem is freed and a new presolved problem is created The routine returns a zero on success and a nonzero if an error occurs int CPXpresolve CPXCENVptr env CPXLPptr lp
93. d to discard solutions that do not meet criteria beyond that of the mixed integer programming formulation Variables are in terms of the original problem if the parameter CPX_PARAM MIPCBREDLP is set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise variables are in terms of the presolved problem The routine returns a zero on success and a nonzero if an error occurs int CPXsetincumbentcallbackfunc CPXENVptr env int CPXPUBLIC incumbentcallback CPXCENVptr env void cbdata int wherefrom void cbhandle double objval double x int isfeas_p int useraction_p void cbhandle CPXENVptr env int void The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXPUBLIC incumbentcallback A pointer to a user written incumbent callback If the callback is set to NULL no callback can be called during optimization cbhandle A pointer to user private data This pointer is passed to the callback The incumbent callback is called when CPLEX has found an integer solution but before this solution replaces the incumbent integer solution ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsetincumbentcallbackfunc Return Value Arguments Variables are in terms of the original problem if the parameter CPX_PARAM MIPCBREDLP is set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise variables are in terms of
94. de files int useraction_p A pointer to an integer indicating the action to be taken on completion of the user callback Table 10 summarizes possible actions Table 10 Actions to be Taken after a User Written Node Callback Value Symbolic Constant Action 0 CPX_CALLBACK_DEFAULT Use ILOG CPLEX selected node 1 CPX_CALLBACK_FAIL Exit optimization 2 CPX_CALLBACK_SET Use user selected node as defined in returned values Example status CPXgetnodecallbackfunc env mynodefunc mydata See Also The example admipexl c CPXgetnodecallbackfunc CPXgetcallbacknodeinfo Advanced MIP Control Interface 166 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsetsolvecallbackfunc CPXsetsolvecallbackfunc Usage Description Return Value Synopsis Arguments CPX int Mixed Integer Users Only The routine CPXsetsolvecallbackfunc sets and modifies the user written callback to be called during MIP optimization to optimize the subproblem The routine returns a zero on success and a nonzero if an error occurs int CPXsetsolvecallbackfunc CPXENVptr env int CPXPUBLIC solvecallback CPXCENVptr env void cbdata int wherefrom void cbhandle int useraction_p void cbhandle ENVptr env El gt The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines CPXPUBLIC solvecallback A pointer to a user written solve callback If the c
95. des routines to translate constraints on the original problem to constraints on the presolved problem so new constraints can be added to the presolved problem As we discuss in Additional Sections it provides a variety of other capabilities When considering mixed integer programs the advanced presolve interface plays a very different role The branch amp cut process needs to be restarted from scratch when the problem is even slightly modified so preserving advanced start information isn t useful The main benefit of the advanced presolve interface for MIPs is that it allows a user to translate decisions made during the branch amp cut process and based on the presolved problem back to the corresponding constraints and variables in the original problem This makes it easier for a user to control the branch amp cut process Details on the advanced MIP callback interface are provided in Chapter 2 Advanced MIP Control Interface Additional Sections The organization of the remainder of this document is as follows Restricting Presolve Reductions discusses the need for restricting the reductions performed by presolve and the mechanisms the advanced presolve interface provides for doing so Manual Control of Presolve discusses routines for manually controlling presolve Modifying a Problem discusses routines for modifying a problem while still retaining presolve information Part II Routines includes detailed descript
96. dnorms turis eee eet a 79 CPXgetbhead o occ carr pees eee A eRe SRS RGA Re RE ERR RR Ee 81 CPXgetbranchcallbackfunc 1 0 etna 82 CPXgetcallbackctype 1 1 tenet tne teens 84 CPXgetcallbackgloballb 2 2 ketene teens 86 CPXgetcallbackglobalub 0 eect tetas 88 CPXgetcallbackincumbent 0 0 0 ete teens 90 CPXgetcallbacklp 2 ketene teens 92 CPXgetcallbacknodeinfo 0 ee ttn ete ae 93 CPXgetcallbacknodeintfeas 1 ketenes 96 CPXgetcallbacknodelb 1 0 0 ect eee 98 CPXgetcallbacknodelp 1 0 0 ett 100 CPXgetcallbacknodeobjval 0 0 0 cect eee tees 102 CPXgetcallbacknodestat 0 eect ete eee 103 CPXgetcallbacknodeub 0 eee teens 104 CPXgetcallbacknodex viii a A oe Rae ee 106 CPXgetcallbackorder 1 0 cee ete eee 108 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 5 TABLE OF CONTENTS CPXgetcallbackpseudocosts oooooococococ tenet eee 110 CPXgetcallbackseqinfo 0 cee tte 112 CPXgetcallbacksosinfO 0 0 cece ete tenes 114 CPXgetcutcallbackfunc esei scrieri cette tees 116 CPXgetdeletenodecallbackfunc 0 cect tenets 117 GPXOCIGNOMNSS sisian arar da a NO 118 CPXgetExactkappa o ooooocoococc eens 119 CPXgetheuristiccallbackfunc 1 0 0 cette eee ee 120 CPXge tdi ees ed re Mea A vega eed eee ty 121 CPXOCUOW sa tio eee od sae A Lea Rawle ee ae Raa eee 122 CPXgetincumbentcallbackfunc oooocococcococe eee 123 G
97. ds certain presolve reductions that make translation from original values to presolved values impossible If the user adds any lazy constraints the user must turn off dual presolve reductions using the CPX_PARAM_REDUCE parameter The user must think carefully about whether constraints added through the cut interface are implied by existing constraints in which case dual presolve reductions may be left on or whether they are not in which case dual reductions are forbidden ILOG Concert Technology users should use the IloCplex LazyConstraintCallbackI when adding lazy constraints and the IloCplex UserCutCallbackI when adding cutting planes Dual reductions and or non linear reductions then will be turned off automatically One scenario that merits special attention is when the user knows a large set of cuts a priori Rather than adding them to the original problem the user may instead wish to add them only when violated The CPLEX advanced MIP control interface provides more than one mechanism for accomplishing this The first and probably most obvious at this point is to install a user callback that checks each cut from the user set at each node adding those that are violated The user can do this either by setting CPX_PARAM_MIPCBREDLP to CPX_OFF to work with the original problem in the cut callback or by using the Advanced Presolve Interface to translate the cuts on the original problem to cuts on the presolved problem and th
98. e Functions before reading this chapter Control callbacks in the ILOG Concert Technology CPLEX Library use original model variables These callbacks are fully documented in the ILOG CPLEX Reference Manual except for the callbacks IloCplex UserCutCallbackI and IloCplex LazyConstraintCallbackI which are documented here Introduction to MIP Callbacks 26 As the reader is no doubt familiar the process of solving a mixed integer programming problem involves exploring a tree of linear programming relaxations CPLEX repeatedly selects a node from the tree solves the LP relaxation at that node attempts to generate cutting planes to cut off the current solution invokes a heuristic to try to find an integer feasible solution close to the current relaxation solution selects a branching variable an integer variable whose value in the current relaxation is fractional and finally places the two nodes that result from branching up or down on the branching variable back into the tree The CPLEX Mixed Integer Optimizer includes methods for each of the important steps listed above node selection cutting planes heuristic branch variable selection incumbent replacement While default CPLEX methods are generally effective and parameters are available to choose alternatives if the defaults are not working for a particular problem there are rare cases where a user may wish to influence or even override CPLEX methods CPLEX provides a callbac
99. e added or not Instead they are there only to strengthen the formulation Note It is an error to use addUserCut for lazy constraints that is constraints whose absence may potentially change the solution of the problem Use addLazyConstraints or equivalently addCut when adding such a constraint When columns are deleted from the extracted model all user cuts are deleted as well and need to be recopied into the user cut pool This method is equivalent to addCuts documented in the ILOG CPLEX Reference Manual TloRangeArray addUserCuts IloRangeArray rng This member function adds a set of user cuts to the invoking IloCplex object Everything said for addUserCut applies to each of the cuts given in array rng void basicPresolve const IloNumVarArray vars IloNumArray redlb 0 IloNumArray redub 0 const IloRangeArray rngs 0 IloBoolArray redundant 0 const This method can be used to compute tighter bounds for the model variables and to detect redundant constraints in the model extracted to the invoking IloCplex object For every variable specified in parameter vars it will return possibly tightened bounds in the corresponding elements of arrays redlb and redub Similarly for every constraint specified in parameter rngs this method will return a boolean indicating whether or not it is redundant in the model in the corresponding element of array redundant void clearLazyConstraints This member function deletes all laz
100. e encourage users to carefully consider whether they can accomplish the same task using the standard CPLEX API before using these advanced routines ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 9 DESIGNATING AND MODIFYING USER WRITTEN CALLBACKS Designating and Modifying User Written Callbacks in MIPs PXsetbranchcallbackfunc PXsetcutcallbackfunc PXsetdeletenodecallbackfunc PXsetincumbentcallbackfunc PXsetnodecallbackfunc 9 96 48 4 4 4 o Cc C G CPXsetheuristiccallbackfunc C C C PXsetsolvecallbackfunc Accessing Current User Written Callbacks in MIPs PXgetbranchcallbackfunc PXgetcutcallbackfunc PXgetdeletenodecallbackfunc PXgetincumbentcallbackfunc PXgetnodecallbackfunc oo o C e C CPXgetheuristiccallbackfunc Cc C C PXgetsolvecallbackfunc Specifying Actions with User Written Callbacks with Routines CPXbranchcallbackbranchbds CPXbranchcallbackbranchconstraints CPXbranchcallbackbranchgeneral CPXcutcallbackadd 9 CPXcutcallbackaddlocal Querying from Within User Written Callbacks in MIPs CPXgetcallbackctype CPXgetcallbackgloballb CPXgetcallbackglobalub IN MIPS 10 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL PXge PXge PXge PXge PXge PXge PXge PXge PXge PXge PXge 96 46 1894 99 9 9 9 9 9 9 E g C C E C C CPXge C C Cc C C PXge Ucal cal cal Ucal tCal tCal Ucal cal DETEC
101. e end of the range of lower bounds to be returned Example status CPXgetcallbacknodelb env cbdata wherefrom Ib 0 cols 1 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 99 CPXgetcallbacknodelp CPXgetcallbacknodelp Usage Mixed Integer Users Only Description The routine CPXgetcallbacknodelp accesses the 1p pointer indicating the currently defined linear programming subproblem LP from within user written callbacks Generally this pointer may be used only in ILOG CPLEX Callable Library query routines such as CPXsolution or CPXgetrows documented in the ILOG CPLEX Reference Manual Q PXgetcallbacknodelp may be called only when its wherefrom argument has one of the following values PX CALLBACK _MIP_BRANCH PX CALLBACK _MIP_CUT PX CALLBACK _MIP_HEURISTIC e E e e PX_CALLBACK_MIP_SOLVE When the wheref rom argument has the value CPX_CALLBACK_MIP_SOLVE the subproblem pointer may also be used in ILOG CPLEX optimization routines Warning Any modification to the subproblem may result in corruption of the problem and of the ILOG CPLEX environment Return Value The routine returns a zero on success and a nonzero if an error occurs A nonzero return value may mean that the requested value is not available Synopsis int CPXgetcallbacknodelp CPXCENVptr env void cbdata int wherefrom CPXLPptr curlp_p Arguments CPXCENVptr env E x lt The pointer
102. e node is integer feasible Custom branching strategies can be implemented by calling the CPLEX function CPXbranchcallbackbranchbds and setting the useraction variable to CPX_CALLBACK_SET Then CPLEX will carry out these branches instead of the CPLEX selected branches Branch variables are in terms of the original problem if the parameter CPX_PARAM MIPCBREDLP is set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise branch variables are in terms of the presolved problem Return Value The callback returns a zero on success and a nonzero if an error occurs Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines sal x lt void cbdata A pointer passed from the optimization routine to the user written callback that identifies the problem being optimized The only purpose of this pointer is to pass it to the callback information routines int wherefrom An integer value indicating where in the optimization this function was called It will have the value CPX_CALLBACK_MIP_BRANCH void cbhandle A pointer to user private data ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 151 CPXsetbranchcallbackfunc int type An integer that indicates the type of branch Table 5 summarizes possible values Table 5 Branch Types Returned from a User Written Branch Callback Symbolic
103. e solutions from the presolved problem creating solutions for the original problem Please note that CPXgetredlp provides the user access to internal CPLEX data structures The presolved problem MUST NOT be modified by the user If the user wishes to manipulate the reduced problem the user should make a copy of it using CPXcloneprob and manipulate the copy instead The advanced presolve interface provides another call that is useful when working directly with the presolved problem either through CPXget redlp or through the advanced MIP callbacks The CPXcrushform call translates a linear expression in the original problem into a linear expression in the presolved problem The most likely use of this routine is to add user cuts to the presolved problem during a mixed integer optimization The advanced presolve interface also provides the reverse operation The CPXuncrushform routine translates a linear expression in the presolved problem into a linear expression in the original problem A limited presolve analysis is performed by CPXbasicpresolve and TloCplex basicPresolve This function determines which rows are redundant and computes strengthened bounds on the variables This information can be used to derive some types of cuts that will tighten the formulation to aid in formulation by pointing out ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 23 MODIFYING A PROBLEM redundancies and to provide upper bounds for variable
104. eckfeas_p A pointer to an integer that indicates whether or not ILOG CPLEX should check the returned integer solution for integer feasibility The solution is checked if checkfeas_p is nonzero When the solution is checked and found to be integer infeasible it is discarded and optimization continues ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsetheuristiccallbackfunc int useraction_p A pointer to an integer to contain the indicator for the action to be taken on completion of the user callback Table 8 summarizes possible values Table 8 Actions to be Taken after a User Written Heuristic Callback Value Symbolic Constant Action 0 CPX_CALLBACK_DEFAULT No solution found 1 CPX_CALLBACK_FAIL Exit optimization 2 CPX_CALLBACK_SET Use user solution as indicated in return values Example status CPXsetheuristiccallbackfunc env myheuristicfunc mydata See Also The example admipex2 c CPXgetheuristiccallbackfunc Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 161 CPXsetincumbentcallbackfunc CPXsetincumbentcallbackfunc Usage Description Return Value Synopsis Arguments Callback Description 162 Mixed Integer Users Only The routine CPXsetincumbentcallbackfunc sets and modifies the user written callback routine to be called when an integer solution has been found but before this solution replaces the incumbent This callback can be use
105. ed from this branch The sequence number may be used to select this node in later calls to the node callback 52 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXbtran CPXbtran Usage Description Return Value Synopsis Arguments Advanced The routine CPXbt ran solves x B y and puts the answer in y B is the basis matrix The routine returns a zero on success and a nonzero if an error occurs int CPXbtran CPXCENVptr env CPXCLPptr lp double y CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines E x lt CPXCLPptr lp A pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double y ILOG An array that holds the right hand side vector on input and the solution vector on output The array must be of length at least equal to the number of rows in the LP problem object CPLEX 8 1 ADVANCED REFERENCE MANUAL 53 CPXcheckax CPXcheckax Usage Description Return Value Synopsis Arguments CPXC Advanced The routine CPXcheckax finds the L norm of Ax b That is this routine checks for numerical roundoff error in the computation of x the resident solution by putting it into that formula and determining which row has the maximum error from zero This routine also returns in one of its arguments the index of the row with the maximum error from zero To
106. ed problem Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetprestat CPXCENVptr env CPXCLPptr lp int prestat_p int pcstat int prstat int ocstat int orstat Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to the original CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int prestat_p A pointer to an integer that will receive the status of the presolved problem associated with LP problem object Ip May be NULL The values for prestat_p are 0 Ip is not presolved or there were no reductions 1 Ip has a presolved problem 2 lp was reduced to an empty problem int pcstat The array where the presolve statuses of the columns are to be returned The array must be of length at least the number of columns in the original problem object May be NULL 128 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetprestat For variable i in the original problem values for pcstat i CPX_PRECOL_LOW CPX_PRECOL_UP CPX_PRECOL_FIX CPX_PRECOL_AGG CPX_PRECOL_OTHE int prstat variable i corresponds to variable pcstat i in the presolved problem variable i is fixed to its lower bound variable i is fixed to its upper bound variable i is fixed to some other value variable i is aggregated
107. ej Pha eee eae 39 GPXAddUSENCUIS zi erriren edges eate een acto a aes yeas ee 40 CPXbasichresolVer decided van behind ede A oe ed eu EA eee ES 41 CPXDINVacO doi A A AA 42 CPXDINVAKOW 562s was A ed ed 43 GPXDINVCOL sitne omai a aid AA e dedicado a ea a aai 44 CPXDINVIOW si ia an id tas 45 CPXbranchcallbackbranchbdS ooocoocccocoo 46 CPXbranchcallbackbranchconstraidtS oooooccocococor tees 48 CPXbranchcallbackbranchgeneral ooccccccccocc 50 GP ADAM nvidia pica e E EN 53 AR exh ee on wen oe aris eta ita land Lanna eee aoa 54 CPXcheckpibsc 2 icici dey es dees dee A eet Se nee edad Pee 55 CPXcopybasednorms ooo 56 CPXcopydnorms sie tte ete eae 58 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL TABLE OF CONTENTS CPXcopypartialbase sf 24 240 coi a 59 CPXCOPYDNOMMS cea siya Ee ae heed ne wee Seda ee ede eee edad 61 CPXcopyprotected e cias tija ad othe Ae AE es ad ao 62 CPXcrushfOrm 232 0022 bos Manette ed a Ad Ga ee eee deed ene 63 CPXCHSNDl e coria whe dnote Gath bdo a cad aioe de atari 65 CPXCIUISNe csc A sagen name eae ae weedy Rear 66 CPXcutcallbackadd 0 0 eee teens 67 CPXcutcallbackaddlocal 0 0 0 cee eee eee 70 CPXP ise see siete edie E daa GA ee eh ead eed le edad 72 CPXdualfarkas issih selec A ek hace bean anes eee Ra RRS 73 CPXfreelazyconstraints 0 0 eee net eee 75 CPXfreepresolve 2 tenet eens 76 CPXireeusercuts vnc cerca bole e A A Se 77 CPX toc Houde ce wears heute rias 78 CPXgetbase
108. en use the presolved cuts in the cut callback Q o gt 2 y 23 0 3 0 7 2 z O U to Another perhaps simpler alternative is to add the cuts or constraints to cut pools before optimization begins This is discussed in Chapter 3 User Cut and Lazy Constraint Pools Branch Selection Callback The next callback we consider is the branch variable selection callback After calling CPXsetbranchcallbackfunc with a pointer to a user callback routine the user routine is called whenever CPLEX makes a branching decision CPLEX indicates which variable has been chosen for branching and allows the user to modify that decision The user may specify the number of children for the current node between 0 and 2 and the set of bounds and or constraints that are modified for each child through calling CPXbranchcallbackbranchbds CPXbranchcallbackbranchconstraints or CPXbranchcallbackbranchgeneral The branch callback routine is called for all viable nodes In particular it will be called for nodes that have zero integer infeasibilities in this case CPLEX will not have chosen a branch variable and the default action will be to discard the node The user can choose to branch from this node and in this way impose additional restrictions on integer solutions A user branch routine may for example call CPXgetcallbacknodeintfeas to identify branching candidates call CPXgetcallbackpseudocosts to obtain pseudo cost ILOG CPL
109. ength at least the number of columns in the LP problem object double slack An array to receive the slack values computed from the x values for the problem object The array must be of length at least the number of rows in the LP problem object Example status CPXslackfromx env lp x slack ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 169 CPXsolwrite CPXsolwrite Usage Advanced Description The routine CPXsolwrite is a generic routine for writing solutions It performs all the calculations needed to produce a solution file but it writes only through functions that the user provides to it so that the user may choose the data representation data selection and file format The user must open the file before calling CPXsolwrite and close the file after calling CPXsolwrite The arguments to CPXsolwrite are functions it calls to write the file CPXsolwrite does not know anything about the file or the type of output being written The argument info the last parameter communicates information to the routines hsection rsectionbeg csectionbeg write_entry and sectionend Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXsolwrite CPXENVptr env CPXLPptr lp void CPXPUBLIC hsection CPXENVptr env CPXLPptr lp void info void CPXPUBLIC rsectionbeg void info void CPXPUBLIC csectionbeg void info void CPXPUBLIC write_entry void
110. ents 160 if it has a better objective value The basis that is saved as part of the incumbent is the optimal basis from the subproblem it may not be a good basis for starting optimization of the fixed problem The integer solution returned to CPLEX is for the original problem if the parameter CPX_PARAM MIPCBREDLP was set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise it is for the presolved problem The callback returns a zero on success and a nonzero if an error occurs CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines ira x void cbdata A pointer passed from the optimization routine to the user written callback to identify the problem being optimized The only purpose of the cbdata pointer is to pass it to the callback information routines int wherefrom An integer value indicating at which point in the optimization this function was called It has the value CPX_CALLBACK_MIP_HEURISTIC for the heuristic callback void cbhandle A pointer to user private data double objval_p A pointer to a variable that on entry contains the optimal objective value of the subproblem and on return contains the objective value of the integer solution found if any double x An array that on entry contains primal solution values for the subproblem and on return contains solution values for the integer solution found if any int ch
111. eous Routines oocoococcnc eee eee eee 13 Concert Technology Library 0 00 cece eee eee 13 Part Concepts csceccedeeawestmeeeaiee sae ceeeeeeeeee 19 Chapter 1 Advanced Presolve Functions 0 0 eee cece eee eens 17 Introduction to Presolve 0 cece cette eee eee eee tenes 17 Restricting Presolve Reductions oooooococonccn eee 19 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 3 TABLE OF CONTENTS Chapter 2 Chapter 3 Part Il Manual Control Of Presolve 0 6 cece eee eee eee eee eee 22 Modifying a Problem 0 0c eee eee eee eee 24 Advanced MIP Control Interface 00 e cece o 25 Introduction to MIP CallbackS oooocooncncnnarar eee eee 26 Heuristic Callback 20 23 205 ernir ii a ee eee o 27 Cut Callback lt lt lt oxcorcccrars bees de eee da Gee cee ee a gad ee eee dele ee ee 28 Branch Selection Callback 2 23 6 csc0c0cc0se ete eee ee a jee 29 Incumbent Callback o lt oooiirs or eee ee ee de eee ene ee eee eee eee ee 30 Node Selection Callback 0 00 cece eee e cece eee e eee eee 31 Solve Callback i 50c000200 ne aprte ee cesar eneae ee eee edd Sees Beek Ree 31 User Cut and Lazy Constraint Pools 000 cece eee eee 33 Adding User Cuts and Lazy Constraints 0 000 c eee eee eee 33 Deleting User Cuts and Lazy ConstraintS oooooococconono eee 36 Routines hiciera OF CPXaddlazyconstraints e se stsenreveds eee at
112. ewhat differently than one might expect If a user wishes to add his own cuts during the branch amp cut process through CPXaddusercuts or CPXcutcallbackadd it may appear necessary to turn off dual reductions to accommodate them However for reasons that are complex and beyond the scope of this discussion dual reductions can be left on The reasons relate to the fact that valid cuts never exclude integer feasible solutions so dual reductions performed for the original problem are still valid after cutting planes are applied However a small set of reductions does need to be turned off Recall that presolve must translate a new constraint on the original problem into a constraint on variables in the presolved problem Most reductions performed by CPLEX presolve replace variables with linear expressions of zero or more other variables plus a constant A few do not These latter reductions make it impossible to perform the translation to the presolved problem Set CPX_PARAM_PRELINEAR to 0 to forbid these latter reductions Restricting the type of presolve reductions will also allow presolve to conclude more about infeasible and or unbounded problems Under the default setting of CPX_PARAM_REDUCE presolve can only conclude that a problem is infeasible and or unbounded If CPX_PARAM_REDUCE is set to CPX_PREREDUCE_PRIMALONLY 1 presolve can conclude that a problem is primal infeasible with return status CPXERR_PRESLV_INF If CPX_P
113. f an error occurs int CPXpivot CPXENVptr env CPXLPptr lp int jenter int jleave int leavestat CPXENVptr env The pointer to the CPLEX environment as returned by the one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int jenter An index indicating the variable to enter the basis The slack or artificial variable for row iis denoted by jenter i 1 jenter must identify a nonbasic variable int jleave An index indicating the variable to leave the basis The slack or artificial variable for row iis denoted by jenter i 1 jleave must identify a basic variable unless jenter denotes a variable with finite upper and lower bounds In that case jleave may be set to jenter to indicate that the variable is moved from its current bound to the other int leavestat An integer indicating the nonbasic status to be assigned to the leaving variable after the basis change This is important for the case where jleave indicates a variable with finite upper and lower bounds as it may become nonbasic at its lower or upper bound status CPXpivot env lp jenter jleave CPX_AT_LOWER ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXpivotin CPXpivotin Usage Description Return Value Synopsis Arguments CPXCI Advanced The routine CPXpivotin forcibly pivots slacks that appear on
114. g specified const int rstat An array of length rent where the gt entry contains the status for slack surplus artificial rindices i For rows other than ranged rows the array element rstat i has the following meaning CPX_AT_LOWER 0 associated slack variable non basic at value 0 0 CPX_BASIC 1 associated slack surplus artificial variable basic For ranged rows the array element rstat i has the following meaning CPX_AT_LOWER o associated slack variable non basic at its lower bound CPX_BASIC 1 associated slack variable basic CPX_AT_UPPER 2 associated slack variable non basic at its upper bound Example status CPXcopypartialbase env lp ccnt colind colstat rent rowind rowstat 60 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXcopypnorms CPXcopypnorms Usage Advanced Description The routine CPXcopypnorms copies the primal steepest edge norms to the specified LP problem object Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXcopypnorms CPXENVptr env CPXLPptr lp double cnorm double rnorm int len Arguments CPXENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double cnorm An array containing values to be used in a subsequent call
115. ger Users Only The routine CPXaddusercuts 1s used to add constraints to the list of constraints that should be added to the LP subproblem of a MIP optimization if they are violated CPLEX handles addition of the constraints and makes sure that all integer solutions satisfy all the constraints The constraints are added to those specified in prior calls to CPXaddusercuts The constraints must be cuts which are implied by the constraint matrix The CPLEX parameter CPX_PARAM_PRELINEAR should be set to CPX_OFF 0 Use CPXfreeusercuts to clear the list of cuts The routine returns a zero on success and a nonzero if an error occurs int CPXaddusercuts CPXCENVptr env CPXLPptr lp int rent int nzcnt const double rhs const char sense const int rmatbeg const int rmatind const double rmatval The arguments of CPXaddusercuts are the same as those of CPXaddrows with the exception that new columns may not be specified so there are no cent and colname arguments and row names may not be specified so there is no rowname argument status CPXaddusercuts env lp cutcnt cutnzcnt cutrhs cutsense cutbeg cutind cutval Example admipex4 c in the advanced examples directory 40 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXbasicpresolve CPXbasicpresolve Usage Advanced Description The routine CPXbasicpresolve performs bound strengthening and detects Return Value Synopsis Argu
116. he routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXuncrushx CPXCENVptr env CPXCLPptr lp double x double prex Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double x An array to receive the primal solution x values for the original problem as computed from primal values of the presolved problem object The array must be of length at least the number of columns in the LP problem object double prex An array that contains primal solution x values for the presolved problem as returned by routines such as CPXgetx and CPXsolution when applied to the presolved problem object The array must be of length at least the number of columns in the presolved problem object Example status CPXuncrushx env lp x prex 180 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXunscaleprob CPXunscaleprob Usage Advanced Description The routine CPXunscaleprob removes any scaling that ILOG CPLEX has applied to the resident problem and its associated data A side effect is that if there is a resident solution any associated factorization is discarded and the solution itself is deactivated meaning that it can no longer be accessed with a call to CPXsolution nor by any other quer
117. hinfo void Example See Also An integer indicating which information is requested Table 3 summarizes possible values Table 3 Information Requested for a User Written SOS Callback Symbolic Constant C Type Meaning CPX_CALLBACK_INFO_SOS_NUM int number of SOSs CPX_CALLBACK_INFO_SOS_TYPE char one of the values in Table 4 CPX_CALLBACK_INFO_SOS_SIZE int size of SOS CPX_CALLBACK_INFO_SOS_IS_FEASIBLE int 1 if SOS is feasible 0 if SOS is not CPX_CALLBACK_INFO_SOS_PRIORITY int priority value of SOS CPX_CALLBACK_INFO_SOS_MEMBER_INDEX int variable index of member member of SOS CPX_CALLBACK_INFO_SOS_MEMBER_REFVAL double reference value weight of this member Table 4 summarizes possible values returned when the type of information requested is the SOS type that is whichinfo CPX_CALLBACK_INFO_SOS_TYPE Table 4 SOS Types Returned when whichinfo CPX_CALLBACK_INFO_SOS_TYPE Symbolic Constant SOS Type CPX_SOS1 type 1 CPX_SOS2 type 2 result_p A generic pointer to a variable of type double int or char representing the value returned by whichinfo The column C Type in Table 3 indicates the type of various values returned by whichinfo status CPXgetcallbacksosinfo env curlp wherefrom 6 4 CPX_CALLBACK_INFO_SOS_IS_FEASIBLE amp isfeasible The example admipex3 c Advanced MI
118. ies or entries exceeding the number of rows CPXpivotin returns an error code Entries of nonslack rows are ignored CPLEX 8 1 ADVANCED REFERENCE MANUAL 141 CPXpivotin int rlen An integer that indicates the number of entries in the array rlist If rlen is negative or greater than the number of rows CPXpivotin returns an error code 142 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXpivotout CPXpivotout Usage Advanced Description The routine CPXpivotout pivots a list of fixed variables out of the resident basis Variables are fixed when the absolute difference between the lower and upper bounds is at most 0e 10 Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXpivotout CPXCENVptr env CPXLPptr lp const int clist int clen Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual const int clist An array of length clen containing the column indices of the variables to be pivoted out of the basis If any of these variables is not fixed CPXpivotout returns an error code int clen An integer that indicates the number of entries in the array clist ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 143 CPXpreaddrows CPXpreaddrows
119. ines to control the branch amp cut process Once CPXpresolve has been called the user can then call CPXgetprestat to obtain information about the reductions performed on the problem This function provides information for each variable in the original problem on whether the variable was fixed and removed aggregated out removed for some other reason or is still present in the reduced problem It also gives information for each row in the original problem on whether it was removed due to redundancy aggregated out or is still present in the reduced problem For those rows and columns that are present in the reduced problem this function provides a mapping from original row column number to row column number in the reduced problem and vice versa 7 gt oO o gt o 2A 0S91d pedueApY Another situation where a user might want to use CPXpresolve is to work directly on the presolved problem By calling CPXgetredlp immediately after CPXpresolve the user can obtain a pointer to the presolved problem For an example of how this might be used the user could call routines CPXcrushx and CPXcrushpi to presolve primal and dual solution vectors call CPXgetredlp to get access to the presolved problem then use CPXcopystart to copy the presolved solutions into the presolved problem then optimize the problem and finally call routines CPXuncrushx and CPXuncrushpi CPXqpuncruspi for QPs to unpresolv
120. info int aflag int num char name char state double vall double val2 double 11 double ul double val3 void CPXPUBLIC sectionend void info void info Arguments CPXENVptr env E x lt The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual 170 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsolwrite void CPXPUBLIC hsection CPXENVptr env CPXLPptr lp void info The function hsection writes a header section in a formatted file Its only arguments are the ILOG CPLEX environment pointer env the problem pointer 1p and the info parameter CPXsolwrite calls this function first It uses the 1p problem pointer to retrieve any information needed by the header section function void CPXPUBLIC rsectionbeg void info The function rsectionbeg writes information at the beginning of the row section of a formatted file It is called after hsection and sectionend Its only argument is the info parameter void CPXPUBLIC csectionbeg void info The function csection writes information at the beginning of the column section of a formatted file It is called after all row entries have been completed Its only argument is the info parameter void CPXPUBLIC write_entry void info int aflag int num char name char
121. ions of each of these routines Note that this current chapter discusses the routines without presenting them in full detail Readers should look to Part II for details on the routines Restricting Presolve Reductions As mentioned in the Introduction to Presolve some presolve reductions are invalidated when a problem is modified The advanced presolve interface therefore allows a user to tell presolve what sort of modifications will be made in the future so presolve can avoid possibly invalid reductions These considerations only apply to linear programs The ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 19 RESTRICTING PRESOLVE REDUCTIONS presolved problem for quadratic programs cannot be modified because of the more complex presolve that is applied to quadratic programs Example Adding Constraints to the First Solution Let us reconsider our proposed example of adding a constraint to a problem after solving it Imagine that you wish to optimize a linear program Primal Dual max xl x x3 min 6y 5y2 st xl x 2x3 lt 6 st yl gt l x2 x3 lt 5 yl y2 gt 1 0 2yl y2 gt 1 xl x2 x3 2 0 yl y2 ys 2 0 Note that the first constraint in the dual y gt is redundant Presolve can use this information about the dual problem and complementary slackness to conclude that variable x can be fixed to 0 and removed from the presolved problem While it may be intuitively obvious from inspection of the primal problem tha
122. ironment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating where the user written callback was called from This parameter must be the value of wnerefrom passed to the user written callback 112 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbackseqinfo int seqnum The sequence number of the node for which information is requested int whichinfo An integer indicating which information is requested For a summary of possible values refer to the table Information Requested for a User Written Node Callback in the description of CPXgetcallbacknodeinfo void result_p A generic pointer to a variable of type double or int representing the value returned by whichinfo The column C Type in the table Information Requested for a User Written Node Callback indicates the type of various values returned by whichinfo ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 113 CPXgetcallbacksosinto CPXgetcallbacksosinfo Usage Description Return Value Synopsis Arguments 114 Mixed Integer Users Only The routine CPXgetcallbacksosinfo accesses information about special ordered sets SOSs during MIP optimization from within user written callbacks This routine may be called only when the value of its wnerefrom argument is one of
123. is true of the various callback support routines CPXcallbackglobalub and so forth If the parameter is set to CPX_OFF 0 all these vectors relate to the variables of the original problem Note that this parameter should not be changed in the middle of an optimization 30 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL NODE SELECTION CALLBACK Node Selection Callback The user can influence the order in which nodes are explored by installing a node selection callback through CPXsetnodecallbackfunc When CPLEX chooses the node to explore next it will call the user callback routine with CPLEX s choice as an argument The callback has the option of modifying this choice Solve Callback The final callback we consider is the solve callback By calling CPXsetsolvecallbackfunc the user instructs CPLEX to call a user function rather than the CPLEX choice dual simplex by default to solve the linear programming relaxations at each node of the tree Advanced example admipex6 c gives an example of how this callback might be used O o gt T y 23 0 30 g a zE O U Note We expect the most common use of this callback will be to craft a customer solution strategy out of the set of available CPLEX algorithms For example a user might create a hybrid strategy that checks for network status calling CPXhybnetopt instead of CPXdualopt when it finds it ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 31 SOLVE CALLB
124. k ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 47 CPXbranchcallbackbranchconstraints CPXbranchcallbackbranchconstraints Description The routine CPXbranchcallbackbranchconstraints specifies the branches to be taken from the current node when the branch is specified by adding one or more constraints to the node problem It may be called only from within a user written branch callback function Constraints are in terms of the original problem if the parameter CPX_PARAM MIPCBREDLP is set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise constraints are in terms of the presolved problem Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXbranchcallbackbranchconstraints CPXCENVptr env void cbdata int wherefrom double nodeest int rent int nzcnt const double rhs const char sense const int rmatbeg const int rmatind const double rmatval void userhandle int seqnum_p Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata A pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value that indicates where the user written callback was called from This parameter must be the value of wherefrom passed to the user written callback double nodeest
125. k mechanism to allow the user to do this If the user installs a callback routine CPLEX calls this routine during the branch amp cut process to allow the user to intervene CPLEX callback functions are thread safe for use in parallel multiple CPU applications Before describing the callback routines we first discuss an important issue related to presolve that the user should be aware of Most of the decisions made within MIP relate to the variables of the problem The heuristic for example finds values for all the variables in the problem that produce a feasible solution Similarly branching chooses a variable on which to branch When considering user callbacks the difficulty that arises is that the user is familiar with the variables in the original problem while the branch amp cut process is performed on the presolved problem Many of the variables in the original problem may have been modified or removed by presolve ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL HEURISTIC CALLBACK CPLEX provides two options for handling the problem of mapping from the original problem to the presolved problem First the user may work directly with the presolved problem and presolved solution vectors This is the default While this option may at first appear unwieldy note that the Advanced Presolve Interface allows the user to map between original variables and presolved variables The downside to this option is that the user has to manually invoke
126. le information 114 changing branch callback 150 branch selection 150 173 cut callback 154 heuristic callback 159 node callback 165 node selection 165 solve callback 167 subproblem optimizer 167 upper or lower bounds 175 checking integer feasibility 160 L_infinity norm 54 55 clearLazyConstraints member function IloCplex class 183 clearUserCuts member function IloCplex class 183 condition number estimate 124 constraint adding with user written callback 28 lazy 28 182 183 redundant 183 constraints adding to LP subproblem of MIP 39 40 CPX CPX_BRANCH_DOWN 109 CPX_BRANCH_GLOBAL 109 CPX_BRANCH_UP 109 200 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPX_CALLBACK_DEFAULT 153 155 161 163 166 168 CPX_CALLBACK_FAIL 153 156 161 163 166 168 CPX_CALLBACK_INFO_MEMBER_INDEX 115 CPX_CALLBACK_INFO_NODE_DEPTH 94 CPX_CALLBACK_INFO_NODE_ESTIMATE 94 CPX_CALLBACK_INFO_NODE_NIINF 94 CPX_CALLBACK_INFO_NODE_OBJVAL 94 CPX_CALLBACK_INFO_NODE_SEQNUM 94 CPX_CALLBACK_INFO_NODE_SIINF 94 CPX_CALLBACK_INFO_NODE_SOS 94 CPX_CALLBACK_INFO_NODE_TYPE 94 CPX_CALLBACK_INFO_NODE_VAR 94 CPX_CALLBACK_INFO_NODES_LEFT 94 CPX_CALLBACK_INFO_SOS_IS_FEASIBLE 115 CPX_CALLBACK_INFO_SOS_MEMBER_REFVAL 115 CPX_CALLBACK_INFO_SOS_NUM114 115 CPX_
127. llback that identifies the problem being optimized The only purpose of this pointer is to pass it to the callback information routines int wherefrom An integer value indicating where in the optimization this function was called It has the value CPX_CALLBACK_MIP_CUT void cbhandle A pointer to user private data int useraction_p A pointer to an integer indicating the action for ILOG CPLEX to take at the completion of the user callback Table 7 summarizes possible actions Table 7 Actions to be Taken After a User Written Cut Callback Value Symbolic Constant Action 0 CPX_CALLBACK_DEFAULT Use cuts as added ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 155 CPXsetcutcallbackfunc Example See Also 156 Table 7 Actions to be Taken After a User Written Cut Callback Value Symbolic Constant Action 1 CPX_CALLBACK_FAIL Exit optimization 2 CPX_CALLBACK_SET Use cuts as added status CPXsetcutcallbackfunc env mycutfunc mydata The example admipex5 c CPXcutcallbackadd CPXgetcutcallbackfunc Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsetdeletenodecallbackfunc CPXsetdeletenodecallbackfunc Usage Description Return Value Synopsis Arguments Callback Description Return Value CPX int Mixed Integer Users Only The routine CPXsetdeletenodecallbackfunc sets and modifies the use
128. llbackfunc This routine does not return a result void CPXgetheuristiccallbackfunc CPXCENVptr env int CPXPUBLIC heuristiccallback_p CPXCENVptr env void cbdata int wherefrom void cbhandle double objval_p double x int checkfeas_p int useraction_p void cbhandle_p CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXPUBLIC heuristiccallback_p The address of the pointer to the current user written heuristic callback If no callback has been set the pointer evaluates to NULL void cbhandle_p The address of a variable to hold the user s private pointer CPXgetheuristiccallbackfunc env amp current_callback amp current_handle CPXsetheuristiccallbackfunc Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetijdiv CPXgetijdiv Usage Advanced Description The routine CPXgetijdiv returns the index of the diverging row that is constraint or column that is variable when one of the ILOG CPLEX simplex optimizers terminates due to a diverging vector This function can be called after an unbounded solution status for a primal simplex call or after an infeasible solution status for a dual simplex call Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXgetijdiv CPXENVptr env CPXLPptr lp int idiv_p int jdiv_p
129. lting residual vector This routine also returns in one of its arguments the index of the basic variable corresponding to this maximum To get the L norm for the scaled problem set the parameter scalrimtype 1 If successful this routine returns the L norm of cg TB If a basic solution does not exist 1 0 is returned double CPXcheckpib CPXCENVptr env CPXCLPptr lp int ijmax_p int scalrimtype CPXCENVptr env El x lt The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines CPXCLPptr lp A pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int ijmax_p A pointer to the row or column with the maximum absolute value in cp TB If i jmax_p corresponds to a row numbered rowindex either a slack row or a ranged row ijmax_p1s 1 rowindex If no solution exists i jmax_p is set to a large integer int scalrimtype An integer that indicates the type of scaling to be applied to the returned L norm When this parameter is equal to 0 zero the returned L norm is unscaled Otherwise the La norm has the same scaling as that applied to the problem currently in memory ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 55 CPXcopybasednorms CPXcopybasednorms Usage Description Return Value Synopsis Arguments 56 Advanced The routine CPXcopybasednorms works in conjunction with the routine CP
130. lve PXqpuncrushpi PXuncrushform PXuncrushpi PXuncrushx Miscellaneous Routines ooo ooo PXcopypartialbase PXdjfrompi PXdualfarkas PXgetray PXmdleave PXqpdjfrompi PXslackfromx PXsolwrite PXstrongbranch PXtightenbds PXunscaleprob Concert Technology Library IloCplex adding the advanced reference methods ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 13 CONCERT TECHNOLOGY LIBRARY 14 2 addLazyConstraint addLazyConstraints addUserCut addUserCuts basicPresolve clearLazyConstraints clearUserCuts importModel with additional parameters TloCplex LazyConstraintCallbackI TloCplex UserCutCallbackI ILOLAZYCONSTRAINTCALLBACK ILOUSERCUTCALLBACK ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL Part Concepts Gn i gt O p o gt 7 ANJOS ld P932ULAPY Advanced Presolve Functions This chapter describes how to use the advanced presolve functions The topics are Introduction to Presolve Restricting Presolve Reductions Manual Control of Presolve Modifying a Problem Introduction to Presolve We begin our discussion of the advanced presolve interface with a quick introduction to presolve Most of the information in this section will be familiar to people who are interested in the advanced interface but we encourage everyone to read through it nonetheless As most CPLEX users know presolve is a process whereby the problem input by the
131. m An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback double x An array to receive the values of the primal variables for the node subproblem This array must be of length at least end begin 1 If successful x 0 through x end begin contain the primal values int begin An integer indicating the beginning of the range of primal variable values for the node subproblem to be returned ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbacknodex int end An integer indicating the end of the range of primal variable values for the node subproblem to be returned Example status CPXgetcallbacknodex env cbdata wherefrom nodex 0 cols 1 See Also Examples admipex1 c admipex3 c and admipex5 c in the advanced examples directory ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 107 CPXgetcallbackorder CPXgetcallbackorder Usage Mixed Integer Users Only Description The routine CPXgetcallbackorder 1s used to get MIP priority order information during MIP optimization from within a user written callback The values are from the original problem if CPX_PARAM_MIPCBREDLP is set to CPX_OFF otherwise they are from the presolved problem This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_
132. ments Example redundant rows CPXbasicpresolve does not create a presolved problem This routine cannot be used for quadratic programs The routine returns a zero on success and a nonzero if an error occurs int CPXbasicpresolve CPXCENVptr env CPXLPptr lp double redlb double redub int rstat CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double redlb An array to receive the strengthened lower bounds The array must be of length at least the number of columns in the LP problem object May be NULL double redub An array to receive the strengthened upper bounds The array must be of length at least the number of columns in the LP problem object May be NULL int rstat An array to receive the status of the row The array must be of length at least the number of rows in the LP problem object May be NULL Values for rstat i O if row i is not redundant 1 if row i is redundant status CPXbasicpresolve env lp reducelb reduceub rowstat ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 41 CPXbinvacol CPXbinvacol Usage Description Return Value Synopsis Advanced The routine CPXbinvacol computes the representation of the j column in terms of the basis In other words it solves Bx
133. mization from within a user written callback The values are from the original problem if cPX_PARAM_MIPCBREDLP is set to CPX_OFF otherwise they are from the presolved problem This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_SOLVE or CPX_CALLBACK_MIP_CUT The routine returns a zero on success and a nonzero if an error occurs int CPXgetcallbackglobalub CPXCENVptr env void cbdata int wherefrom double ub int begin int end CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback double ub An array to receive the values of the global upper bound values This array must be of length at least end begin 1 If successful ub 0 through ub end begin contain the global upper bound values int begin An integer indicating the beginning of the range of upper bound values to be returned ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallb
134. mputes information for selecting a branching variable in an integer programming branch amp cut search To describe this routine let s assume that an LP has been solved and that the optimal solution is resident Let good1ist be the list of variable indices for this problem and goodlen be the length of that list Then good1ist gives rise to 2 goodlen different LPs in which each of the listed variables in turn is fixed to the greatest integer value less than or equal to its value in the current optimal solution and then each variable is fixed to the least integer value greater than or equal to its value in the current optimal solution CPXstrongbranch performs at most it 1im dual steepest edge iterations on each of these 2 goodlen LPs starting from the current optimal solution of the base LP The values that these iterations yield are placed in the arrays downpen for the downward fix and uppen for the upward fix Setting CPX_PARAM_DGRADIENT to 2 may give more informative values for the arguments downpen and uppen for a given number of iterations itlim A user might use other routines of the ILOG CPLEX Callable Library directly to build a function that computes the same values as CPXst rongbranch However CPXstrongbranch should be faster because it takes advantage of direct access to internal ILOG CPLEX data structures The routine returns a zero on success and a nonzero if an error occurs int CPXstrongbranch CPXCENVp
135. n about SOS from callbacks 114 kappa 124 L_ infinity norms 55 lower bound values for subproblem from user written call back 98 lower bound values from user written callback 86 members of SOS 114 node callback information 112 objective offset 126 objective value for subproblem from user written callback 102 objective value of LP subproblem 94 pointer to MIP problem used by callback function 92 pointer to subproblem 100 position of basic variable in basis header 122 primal variable values for subproblem from user written callback 106 ILOG CPLEX 8 1 Index priority order from user written callback 108 pseudo cost values from user written callback 110 right hand side vector 78 subproblem optimization status from user written callback 103 upper bound values for subproblem from user written call back 104 upper bound values from user written callback 88 adding lazy constraints 182 user cuts 182 addLazyConstraint member function IloCplex class 182 addLazyConstraints member function IloCplex class 182 addUserCut member function IloCplex class 182 addUserCuts member function IloCplex class 183 basicPresolve member function IloCplex class 183 basis 42 53 56 78 79 81 181 bounds on variables 175 copying partial 59 deleting dual steepest edge norms 136 deleting primal steepest edge norms 137 inverse 44 45 ADVANCED REFERENCE MANUAL 199 INDEX pivoting fixed variables out of 143 pivoting slacks i
136. n the presolved problem for it to be branched on The user should use the CPXgetcallbacknodeintfeas routine from the Advanced Presolve Interface to find branching candidates those for which CPXgetcallbacknodeintfeas returns CPX_INTEGER_INFEASIBLE The CPXcopyprotected routine can be used to prevent presolve from removing specific variables from the presolved problem While restricting branching may appear to limit your ability to solve a problem in fact a problem can always be solved to optimality by branching only on the variables of the presolved problem Incumbent Callback The incumbent callback is used to reject integer feasible solutions that do not meet additional restrictions the user may wish to impose The user callback routine will be called each time a new incumbent solution has been found including when solutions are provided by the user s heuristic callback routine The user callback routine is called with the new solution as input The callback function should return a value that indicates whether or not the new solution should replace the incumbent solution As with other MIP control callback routines the CPX_PARAM_MIPCBREDLP parameter influences the arguments to the user callback routine If this parameter is set to its default value of CPX_ON 1 the solution vector that is input to the callback is a presolved vector It contains one value for each variable in the presolved problem The same
137. nd column indices are specified in the same call The routine returns a zero on success and a nonzero if an error occurs int CPXgetijrow CPXCENVptr env CPXCLPptr lp int i Int jy int row_p ENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp The pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int i int An integer specifying the index of a basic row CPXgetijrow must find the position of this basic row in the basis header A negative value in this argument indicates to CPXgetijrow not to seek a basic row j An integer specifying the index of a basic column CPXgetijrow must find the position of this basic column in the basis header A negative value in this argument indicates to CPXgetijrow not to seek a basic column int row_p 122 ILOG A pointer to an integer indicating the position in the basis header of the row i or column j If CPXgetijrow encounters an error and if row_p is not NULL row_p is set to 1 CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetincumbentcallbackfunc CPXgetincumbentcallbackfunc Usage Description Return Value Synopsis Arguments Example See Also Mixed Integer Users Only The routine CPXgetincumbentcallbackfunc accesses the user written callback to be called by CPLEX during MIP optimization after an integer
138. node selection 165 solve callback 167 subproblem 100 subproblem optimizer 159 167 upper or lower bounds 175 N node callback to change 125 165 exploring 125 165 fathomed 152 files 112 166 index 166 numbering conventions 166 selection 125 165 sequence number of 94 166 solving linear problem at 18 viable 27 node selection callback 31 node subproblem determining integer feasibility of variables 96 nonbasic 61 norms dual steepest edge 58 79 118 136 L_ infinity 55 primal steepest edge 61 127 137 numbering conventions cuts 154 dual steepest edge norms 58 118 nodes 166 special ordered sets SOS 114 O objective function changing coefficients of problem object and presolved problem 145 objective offset 126 INDEX objective value accessing for subproblem from user written callback 102 optimality resident solution 173 subproblem solutions and 159 optimization status accessing for subproblem from user written callback 103 optimizing subproblem 135 167 P pivoting fixed variables out of basis 143 slacks into basis 141 pointer 134 presolve basic 183 gathering information about 23 interface 22 limited 23 performing 146 process 17 process for MIP 18 protecting variables during 22 restricting dual reductions 20 restricting primal reductions 21 presolve member function IloCplex class 185 presolved problem 126 accessing pointer to 134 adding constraints to 19 adding rows 144 and branch amp c
139. nstraints that is not implied by the constraints in the constraint matrix The parameter CPX_PARAM_PRELINEAR must be set to 0 if the constraints to be added are in terms of the original problem and the constraints are valid cutting planes The routine returns a zero on success and a nonzero if an error occurs int CPXsetcutcallbackfunc CPXENVptr env int callback CPXCENVptr xenv void cbdata int wherefrom void cbhandle int useraction_p void cbhandle ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsetcutcallbackfunc Arguments Callback Description Return Value Arguments CPXENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines int callback The pointer to the current user written cut callback If no callback has been set the pointer evaluates to NULL void cbhandle A pointer to user private data This pointer is passed to the user written cut callback ILOG CPLEX calls the cut callback when the LP subproblem for a node has an optimal solution with objective value below the cutoff and is integer infeasible The callback returns a zero on success and a nonzero if an error occurs CPXCENVptr xenv The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata A pointer passed from the optimization routine to the user written ca
140. nter to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual const double norm An array containing values to be used in a subsequent call to CPXdualopt with a setting of CPX_PARAM_DGRADIENT different from 1 one as the initial values for the dual steepest edge norms of the corresponding basic variables specified in head The array must be of length at least equal to the value of the argument len If any indices in head are not basic the corresponding values in norm are ignored const int head An array containing the indices of the basic variables for which norms have been specified in norm The array must be of length at least equal to the value of the argument len int len An integer that indicates the number of entries in norm and head CPXcopypnorms CPXgetdnorms ILOG CPLEX 8 1 ADVA NCED REFERENCE MANUAL CPXcopypartialbase CPXcopypartialbase Description Return Value Synopsis Arguments The routine CPXcopypartialbase is used to copy a a partial basis into an LP problem object Basis statuses do not need to be specified for every variable or slack surplus artificial variable If the status of a variable is not specified it is made non basic at lower bound if the lower bound is finite otherwise non basic at upper bound if the upper bound is finite otherwise non basic at 0 0 If the status of a slack surplus artificial v
141. nto 141 replacing one variable with another 140 starting from advanced 18 basis header 81 122 bounds strengthening 41 tighter 183 branch amp cut 27 67 70 116 154 173 controlling process with presolve 19 branch amp cut tree solving linear problem at each node 18 branch callback 82 150 fathoming nodes 152 node depth 94 type of 94 branch selection changing 82 150 173 from current node 46 48 50 strong 173 branch variable selection callback 29 branch types of 94 152 C callback accessing ctypes 84 accessing incumbent values 90 accessing lower bound values 86 accessing lower bound values for subproblem 98 accessing MIP problem in use 92 accessing objective value for subproblem 102 accessing primal variable values for subproblem 106 accessing priority order 108 accessing pseudo cost values 110 accessing subproblem optimization status 103 accessing upper bound values 88 accessing upper bound values for subproblem 104 adding cuts 67 70 adding cuts to subproblems 116 154 after branch selection 82 150 before replacing incumbent 123 branch 82 150 branch variable selection 29 cut 28 defining with macro 188 189 deleting a node 117 heuristic 27 120 159 incumbent 30 infeasibility and 120 159 lazy constraint 186 188 node selection 31 optimizing subproblem 135 167 set node for deletion 157 solve 31 user cut 187 189 user written 67 70 82 100 114 116 117 120 135 150 154 157 159 162 167 variab
142. o the user callback routine If this parameter is set to its default value of CPX_ON 1 the solution vector returned to the callback and any feasible solutions returned by the callback are presolved vectors They contain one value for each variable in the presolved problem The same is true of the various callback support routines CPXgetcallbackgloballb etc If the parameter is set to CPX_OFF 0 all these vectors relate to variables of the original problem Note that this parameter should not be changed in the middle of an optimization The user should be aware that the branch amp cut process works with the presolved problem so the code will incur some cost when translating from presolved to original values This cost is usually small but can sometimes be significant We should also note that if a user wishes to solve linear programs as part of a heuristic callback the user must make a copy of the node LP for example using CPXcloneprob The user should not modify the CPLEX node LP The next example we consider is the user cut callback routine The user calls CPXsetcutcallbackfunc to set a cut callback and the user s callback routine is called at every viable node of the branch amp cut tree We refer the reader to admipex5 c fora detailed example A likely sequence of events once the user callback function is called is as follows First the routine calls cPXgetcallbacknodex to get the relaxation solution for the c
143. o the original problem This information is hidden within the user s problem in the CPLEX LP problem object for Callable Library users and was inaccessible to the user in CPLEX releases prior to 7 0 The presolve process for a mixed integer program is similar but has a few important differences First the actual presolve reductions differ Integrality restrictions allow CPLEX to perform several classes of reductions that are invalid for continuous variables A second difference is that the MIP solution process involves a series of linear program solutions In the MIP branch amp cut tree a linear program is solved at each node MIP presolve is performed once at the beginning of the optimization unless the CPX_PARAM_RELAXPREIND parameter is set in which case the root relaxation is presolved a second time and all of the node relaxation solutions use the presolved problem Again presolve stores the presolved problem and the information required to convert a presolved solution to a solution for the original problem within the LP problem object Again this information was inaccessible to the user in CPLEX releases prior to version 7 0 A Proposed Example Now consider an application where the user wishes to solve a linear program using the simplex method then modify the problem slightly and solve the modified problem As an example let s say a user wishes to add a few new constraints to a problem based on the results of the first
144. of its wherefrom argument is CPX_CALLBACK_MIP_CUT The cut may be for the original problem if the parameter CPX_PARAM_MIPCBREDLP was set to CPX_OFF before the call to CPXmipopt that calls the callback Otherwise the cut is used on the presolved problem The routine returns a zero on success and a nonzero if an error occurs int CPXcutcallbackaddclocal CPXCENVptr env void cbdata int wherefrom int nzcnt double rhs int sense const int cutind const double cutval CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines sal x lt void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value that indicates where the user written callback was called from This parameter must be the value of wherefrom passed to the user written callback int nzcnt An integer value that indicates the number of coefficients in the cut or equivalently the length of the arrays cutind and cutval double rhs A double value that indicates the value of the right hand side of the cut ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXcutcallbackaddlocal int sense An integer value that indicates the sense of the cut const int cutind An array containing the column indices of cut coefficients const double cutval An array containing the v
145. of length at least goodlen The entries in goodlist must all be indices of current basic variables Moreover these indices must all be indices of original model variables that is they must all take values smaller than the number of columns in the model as returned by CPXgetnumcol s Negative indices and indices bigger than or equal to CPXgetnumcols result in an error int goodlen An integer indicating the number of entries in goodlist If goodlen lt 0 an error is returned double downratio An array of type double that must be of length at least goodlen 138 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXmdleave For a given j goodlist i downratio i has the following meaning Let x be the name of the basic variable with index j and suppose that x is fixed to some value 1 lt t Ina subsequent call to CPXdualopt the leaving variable in the first iteration of this call is uniquely determined It must be x There are then two possibilities Either an entering variable is determined or it is concluded in the first iteration that the changed model is dual unbounded primal infeasible In the latter case downratio i is set equal to a large positive value this number is system dependent but is usually 0 75 In the former case where r is the value of the objective function after this one iteration downratio i is determined by Irl t t downratio i double upratio An array of type
146. of the CPXopenCPLI routines El x lt CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double cnorm An atray containing the primal steepest edge norms for the normal column variables The array must be of length at least equal to the number of columns in the LP problem object double rnorm An array containing the primal steepest edge norms for ranged variables and slacks The array must be of length at least equal to the number of rows in the LP problem object int len_p A pointer to the number of entries in the array cnorm When this routine is called len_pis equal to the number of columns in the LP problem object when optimization occurred The routine CPXcopypnorms needs the value 1en_p There is no comparable argument in this routine for rnorm If the rows of the problem have changed since the norms were computed they are generally no longer valid However if columns have been deleted or if columns have been added the norms for all remaining columns present before the deletions or additions remain valid See Also CPXcopypnorms ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 127 CPXgetprestat CPXgetprestat Usage Advanced Description The routine CPXgetprestat is used to get presolve status information for the columns and rows of the presolved problem in the original problem and of the original problem in the presolv
147. of the basic variables in the order specified by head The array must be of length at least equal to number of rows in the LP problem object CPLEX 8 1 ADVANCED REFERENCE MANUAL 81 CPXgetbranchcallbackfunc CPXgetbranchcallbackfunc Usage Description Return Value Synopsis Arguments Example 82 Mixed Integer Users Only The routine CPXgetbranchcallbackfunc accesses the user written callback routine to be called during MIP optimization after a branch has been selected but before the branch is carried out ILOG CPLEX uses the callback routine to change its branch selection For documentation of callback arguments see the routine CPXsetbranchcallbackfunc This routine does not return a result void CPXgetbranchcallbackfunc CPXCENVptr env int CPXPUBLIC branchcallback CPXCENVptr env void cbdata int wherefrom void cbhandle int type int sos int nodecnt int bdcnt double nodeest int nodebeg int indices char lu int bd int useraction_p void cbhandle CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines int CPXPUBLIC branchcallback The address of the pointer to the current user written branch callback If no callback has been set the returned pointer evaluates to NULL void cbhandle The address of a variable to hold the user s private pointer CPXgetbranchcallbackfunc env
148. ointer evaluates to NULL void cbhandle_p The address of a variable to hold the user s private pointer CPXgetsolvecallbackfunc env amp current_callback amp current_cbdata CPXgetcallbacknodelp CPXsetsolvecallbackfunc Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 135 CPXkilldnorms CPXkilldnorms Usage Advanced Description The routine CPXkilldnorms deletes any dual steepest edge norms that have been retained relative to an active basis If the user believes that the values of these norms may be significantly in error and the setting of the CPX_PARAM_DPRI IND parameter is CPX_DPRIIND_STE EP or CPX_DPRIIND_FULLSTE EP calling CPXkilldnorms means that fresh dual steepest edge norms will be computed on the next call to CPXdualopt Synopsis void CPXkilldnorms CPXLPptr lp Arguments CPXLPptr lp The pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the 136 ILOG CPLEX 8 1 CPLEX Reference Manual ADVANCED REFERENCE MANUAL CPXkillpnorms CPXkillpnorms Usage Advanced Example The routine CPXkillpnorms deletes any primal steepest edge norms that have been retained relative to an active basis If the user believes that the values of these norms may be significantly in error and the setting of the CPX_PARAM_PPRIIND parameter is CPX_PPRIIND_STEEP calling CPXkillpnorms means that fresh dual
149. on Return Value Synopsis Arguments Advanced The routine CPXtightenbds changes the upper or lower bounds on a set of variables in a problem Several bounds can be changed at once Each bound is specified by the index of the variable associated with it The value of a variable can be fixed at one value by setting both the upper and lower bounds to the same value In contrast to the ILOG CPLEX Callable Library routine CPXchgbds also used to change bounds CPXtightenbds preserves more of the internal ILOG CPLEX data structures so it is more efficient for re optimization particularly when changes are made to bounds on basic variables The routine returns a zero on success and a nonzero if an error occurs int CPXtightenbds CPXCENVptr env CPXLPptr lp int cnt const int indices const char lu const double bd CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines E gt CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual int cnt An integer indicating the total number of bounds to change That is cnt specifies the length of the arrays indices lu and bd const int indices An array containing the numerical indices of the columns corresponding to the variables for which bounds will be changed The allocated length of the array is cnt Column j of the constraint ma
150. onstraints from the branch callback They are CPXbranchcallbackbranchconstraints where the branch is created by adding one or more constraints to a problem and CPXbranchcallbackbranchgeneral where the branch is created by adding one or more constraints and changing one or more variable bounds Local Cuts The function CPXcutcallbackaddlocal is called from the cut callback to add cuts that apply to the subtree rooted at the current node This contrasts with the function CPXcutcallbackadd which is used to add global cuts that is cuts that are valid at all nodes Associating User Data with Nodes The delete node callback is called when a node is removed from the branch amp bound tree either by branching on it or by fathoming it It can be used to free user data that was associated with a node The routine CPXsetdeletenodecallbackfunc instructs CPLEX to use the specified user written callback routine whenever a node is deleted The routine CPXsetdeletenodecallbackfunc accesses the user written routine currently being used 192 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPLEX 7 1 ADVANCED FEATURES RELEASE NOTES Obtaining Node Information by Sequence Number The new routine CPXgetcallbackseqinfo is used to query branch amp cut nodes during the node callback The node sequence number is used to identify the node instead of the node index number Information on the Current Node The routine CPXgetcallbacknodeinfo
151. or callbacks The implementation of method main must be provided by the user in parentheses following the macro invocation ILOUSERCUTCALLBACKn name implementation of the callback For the implementation of the callback methods from class IloCplex UserCutCallbackI and its parent classes can be used You are not obliged to use this macro to define callbacks When the macro seems too restrictive for your purposes we recommend that you define a callback class directly Since the argument name is used to name the callback class it is not possible to use the same name for several callback definitions lt ilcplex ilocplex h gt lt ilcplex ilocplex h gt TloCplex UserCutCallbackI ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 189 190 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL Advanced Features Release Notes This appendix includes 4 CPLEX 8 1 Advanced Features Release Notes CPLEX 8 0 Advanced Features Release Notes CPLEX 7 1 Advanced Features Release Notes CPLEX 6 6 to 7 0 Advanced Features Release Notes CPLEX 8 1 Advanced Features Release Notes No conversion steps pertaining to advanced features are necessary in order to move from CPLEX 8 0 to CPLEX 8 1 This is also the case with the standard features of CPLEX as documented in the CPLEX 8 1 Release Notes CPLEX 8 0 Advanced Features Release Notes ILOG CPLEX 8 0 offers a number of new routines and extensions
152. out variable i is deleted or merged for some other reason The array where the presolve statuses of the rows are to be returned The array must be of length at least the number of rows in the original problem object May be NULL For row i in the original problem values for prstat i CPX_PREROW_RED CPX_PREROW_AGG CPX_PREROW_OTHE int ocstat row i corresponds to row prstat i in the original problem if row i is redundant if row i is used for aggregation if row i is deleted for some other reason The array where the presolve statuses of the columns of the presolved problem are to be returned The array must be of length at least the number of columns in the presolved problem object May be NULL For variable i in the presolved problem values for ocstat i 20 variable i in the presolved problem corresponds to variable ocstat i in the original problem 1 variable i corresponds to a linear combination of some variables in the original problem int orstat The array where the presolve statuses of the rows of the presolved problem are to be returned The array must be of length at least the number of rows in the presolved problem object May be NULL For row i in the original problem values for orstat i 20 __ if row i in the presolved problem corresponds to row orstat i in the original problem 1 if row i is created by for example merging two rows in the original problem ILO
153. outine returns a zero on success and a nonzero if an error occurs Synopsis int CPXsetbranchcallbackfunc CPXENVptr env int CPXPUBLIC branchcallback CPXCENVptr env void cbdata int wherefrom void cbhandle int type int sos int nodecnt int bdcnt double nodeest int nodebeg int indices char lu int bd int useraction_p void cbhandle Arguments CPXENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines int CPXPUBLIC branchcallback A pointer to a user written branch callback If the callback is set to NULL no callback can be called during optimization void cbhandle A pointer to user private data This pointer is passed to the callback 150 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsetbranchcallbackfunc Callback Description The call to the branch callback occurs after a branch has been selected but before the branch is carried out This function is written by the user On entry to the callback the ILOG CPLEX selected branch is defined in the arguments The arguments to the callback specify a list of changes to make to the bounds of variables when child nodes are created One two or zero child nodes can be created so one two or zero lists of changes are specified in the arguments The first branch specified is considered first The callback is called with zero lists of bound changes when the solution at th
154. p Arguments CPXCENVptr env E x lt The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines CPXCLPptr lp A pointer to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual 124 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetnodecallbackfunc CPXgetnodecallbackfunc Usage Description Return Value Synopsis Arguments Example See Also CPXCI Mixed Integer Users Only The routine CPXgetnodecallbackfunc accesses the user written callback to be called during MIP optimization after ILOG CPLEX has selected a node to explore but before this exploration is carried out The callback routine can change the node selected by ILOG CPLEX to a node selected by the user For documentation of callback arguments see the routine CPXsetnodecallbackfunc This routine does not return a result void CPXgetnodecallbackfunc CPXCENVptr env int CPXPUBLIC nodecallback_p CPXCENVptr env void cbdata int wherefrom void cbhandle int nodeindex_p int useraction_p void cbhandle_p int void ILOG ENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines E x lt CPXPUBLIC nodecallback_p The address of the pointer to the current user written node callback If no callback has been set the pointer will evaluate to NULL cbhandle_p
155. point out a few of the subtleties associated with using CPXcopystart to start an optimization from an advanced presolved solution This routine only creates a presolved solution vector if the presolved problem is already present either because the user called CPXpresolve or because the user turned off some presolve reductions through CPX_PARAM_REDUCE and then solved a problem The earlier sequence would not have started from an advanced solution if CPXcopystart had been called before CPXpresolve Another important detail about CPXcopystart is that it crushes primal and or dual solutions not bases It then uses the crushed solutions to choose a starting basis If you have a basis you need to compute optimal primal and dual solutions using 22 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL MANUAL CONTROL OF PRESOLVE CPXcopybase and then CPXprimopt extract them and then call CPXcopystart with them to use the corresponding advanced solution In general starting with both a primal and dual solution is preferable to starting with one or the other One other thing to note about CPXcopystart is that the primal and dual slack slack and dj arguments are optional The routine will compute slacks values if none are provided Another situation where a user might want to use CPXpresolve is if the user wishes to gather information about the presolve possibly in preparation for using the advanced MIP callback rout
156. r written callback routine to be called during MIP optimization when a node is to be deleted Nodes are deleted when a branch is carried out from that node when the node relaxation is infeasible or when the node relaxation objective value is worse than the cutoff The routine returns a zero on success and a nonzero if an error occurs int CPXsetdeletenodecallbackfunc CPXENVptr env int CPXPUBLIC deletenodecallback CPXCENVptr env void cbdata int wherefronm int seqnum void handle void cbhandle ENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXPUBLIC deletenodecallback A pointer to a user written branch callback If the callback is set to NULL no callback can be called during optimization void cbhandle A pointer to user private data This pointer is passed to the callback The call to the delete node callback routine occurs during MIP optimization when a node is to be deleted Nodes are deleted when a branch is carried out from that node when the node relaxation is infeasible or when the node relaxation objective value is worse than the cutoff The main purpose of the callback is to provide an opportunity to free any user data associated with the node thus preventing memory leaks The callback returns a zero on success and a nonzero if an error occurs ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 157 CPXsetdeletenodecallbackfunc
157. reated as lazy constraints When the node status is unbounded then the function CPXgetcallbacknodex returns a ray that can be used to decide which lazy constraints need to be added to the subproblem This routine may be called only when the value of the wherefrom argument is CPX_CALLBACK_MIP_CUT The routine returns a zero on success and a nonzero if an error occurs int CPXgetcallbacknodestat CPXCENVptr env void cbdata int wherefrom int nodestat_p CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLI routines na pS void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback int nodestat_p A pointer to an integer where the node subproblem optimization status is to be returned The values of nodestat_p may be CPX_OPTIMAL or CPX_UNBOUNDED status CPXgetcallbacknodestat env cbhdata wherefrom amp nodestatus ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 103 CPXgetcallbacknodeub CPXgetcallbacknodeub Usage Description Return Value Synopsis Arguments 104 Mixed Integer Users Only The routine CPXgetcallbacknodeub is used to get the upper bound values for the subproblem a
158. redlp CPXCENVptr env CPXCLPptr lp CPXCLPptr redlp_p Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual CPXCLPptr redlp_p A pointer to receive the problem object pointer that results when presolve has been applied to the LP problem object Example status CPXgetredlp env lp amp reducelp 134 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetsolvecallbacktunc CPXgetsolvecallbackfunc Usage Description Return Value Synopsis Arguments Example See Also Mixed Integer Users Only The routine CPXgetsolvecallbackfunc accesses the user written callback to be called during MIP optimization to optimize the subproblem For documentation of callback arguments see the routine CPXsetsolvecallbackfunc This routine does not return a result void CPXgetsolvecallbackfunc CPXCENVptr env int CPXPUBLIC solvecallback_p CPXCENVptr env void cbdata int wherefrom void cbhandle int useraction_p void cbhandle_p CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines lkal x lt int CPXPUBLIC solvecallback_p The address of the pointer to the current user written solve callback If no callback has been set the p
159. refrom passed to the user written callback int nzcnt An integer value that indicates the number of coefficients in the cut or equivalently the length of the arrays cutind and cutval double rhs A double value that indicates the value of the right hand side of the cut ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 67 CPXcutcallbackadd int sense An integer value that indicates the sense of the cut const int cutind An array containing the column indices of cut coefficients const double cutval An array containing the values of cut coefficients Example status CPXcutcallbackadd env cbdata wherefrom mynzcnt myrhs g P mycutind mycutval See Also The example admipex5 c CPXgetcutcallbackfunc CPXsetcutcallbackfunc 68 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXcutcallbackadd ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 69 CPXcutcallbackaddlocal CPXcutcallbackaddlocal Usage Description Return Value Synopsis Arguments 70 Mixed Integer Users Only The routine CPXcutcallbackaddlocal adds local cuts during MIP branch amp cut A local cut is one that applies to the current nodes and the subtree rooted at this node Global cuts that is cuts that apply throughout the branch amp cut tree are added with the routine CPXcutcallbackadd This routine may be called only from within user written cut callbacks thus it may be called only when the value
160. rol Interface O o gt 2e y 23 0 3 0 7 2 z O U This chapter describes the CPLEX 8 1 advanced MIP control interface It includes sections on Introduction to MIP Callbacks Heuristic Callback Cut Callback Branch Selection Callback Incumbent Callback Node Selection Callback Solve Callback 9 gt gt 4 4 These callbacks allow sophisticated users to control the details of the branch amp cut process Specifically users can choose the next node to explore choose the branching variable add their own cutting planes place additional restrictions on integer solutions or insert their own heuristic solutions These functions are meant for situations where other tactics to improve CPLEX s performance on a hard MIP problem such as non default parameter settings or priority orders have failed We refer the reader to the section on Troubleshooting MIP Performance Problems in the LOG CPLEX User s Manual for more information on MIP parameters and priority orders ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 25 INTRODUCTION TO MIP CALLBACKS Users of the advanced MIP control interface can work with the variables of the presolved problem or by following a few simple rules can instead work with the variables of the original problem Tip The advanced MIP control interface relies heavily on the advanced presolve capabilities We suggest that the reader become familiar with Chapter 1 Advanced Presolv
161. rray rngs TloRangeArray lazy 0 TloRangeArray cuts 0 const This method is an extension of the importModel method documented in the Reference Manual but has two extra parameters lazy and cuts If a non zero array handle is passed as parameter lazy it will be filled with IloRange objects corresponding to all lazy constraints that are found in the model file Similarly if a non zero array handle is passed as parameter cuts it will be filled with I1oRange objects corresponding to all user cuts that are found in the model file The rest of the parameters are exactly the same as the parameters of the corresponding importModel method documented in the ILOG CPLEX Reference Manual void importModel IloModel amp m const char filename TloObjective amp obj IloNumVarArray vars IloRangeArray rngs ITloSoSilArray sosl TloSOS2Array sos2 IloRangeArray lazy 0 TloRangeArray cuts 0 const This method is an extension of the importModel method documented in the Reference Manual but has two extra parameters lazy and cuts If a non zero array handle is passed as parameter lazy it will be filled with IloRange objects corresponding to all lazy constraints that are found in the model file Similarly if a non zero array handle is passed as parameter cuts it will be filled with I1oRange objects corresponding to all user cuts that are found in the model file The rest of the parameters are exactly the same as the parameters of the corre
162. s Arguments Example Advanced The routine CPXuncrushpi uncrushes a dual solution for the presolved problem to a dual solution for the original problem This routine is for linear programs Use CPXqpuncrushpi for quadratic programs The routine returns a zero on success and a nonzero if an error occurs int CPXuncrushpi CPXCENVptr env CPXCLPptr lp double pi const double prepi CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double pi An array to receive dual solution pi values for the original problem as computed from the dual values of the presolved problem object The array must be of length at least the number of rows in the LP problem object const double prepi An array that contains dual solution pi values for the presolved problem as returned by routines such as CPXgetpi and CPXsolution when applied to the presolved problem object The array must be of length at least the number of rows in the presolved problem object status CPXuncrushpi env lp pi prepi ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 179 CPXuncrushx CPXuncrushx Usage Advanced Description The routine CPXuncrushx uncrushes a solution for the presolved problem to the solution for the original problem Return Value T
163. s are equivalent to but sometimes different from those input by the user The dual variables associated with the added rows may take different values than those the user might expect If a user makes a problem modification that is not consistent with the setting of CPX_PARAM_REDUCE the presolved problem is discarded and presolve is reinvoked at the next optimization call Similarly CPLEX discards the presolved problem if the user modifies a variable or constraint that presolve had previously removed from the problem You can use CPXpreaddrows or CPXprechgob3 to make sure that this will not happen Note that CPXpreaddrows also permits changes to the bounds of the presolved problem If the nature of the procedure dictates a real need to modify the variables that presolve removed you can use the CPXcopyprotected routine to instruct CPLEX not to remove those variables from the problem Instead of changing the bounds on the presolved problem consider changing the bounds on the original problem CPLEX will discard the presolved problem but calling CPXpresolve will cause CPLEX to apply presolve to the modified problem with the added benefit of reductions based on the latest problem modifications Then use CPXcrushx CPXcrushpi and CPXcopystart to provide an advanced start for the problem after presolve has been applied on the modified problem 24 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL Advanced MIP Cont
164. s The constraints are added to those specified in prior calls to CPXaddlazyconstraints Lazy constraints are constraints not specified in the constraint matrix of the MIP problem but that must be not be violated in a solution Using lazy constraints makes sense when there are a large number of constraints that must be satisfied at a solution but are unlikely to be violated if they are left out The CPLEX parameter CPX_PARAM_REDUCE should be set to CPX_PREREDUCE_NOPRIMALORDUAL 0 or to CPX_PREREDUCE_PRIMALONLY 1 in order to turn off dual reductions Use CPXfreelazyconstraints to clear the list of lazy constraints The routine returns a zero on success and a nonzero if an error occurs int CPXaddlazyconstraints CPXCENVptr env CPXLPptr lp int rent int nzcnt const double rhs const char sense const int rmatbeg const int rmatind const double rmatval The arguments of CPXaddlazyconstraints are the same as those of CPXaddrows with the exception that new columns may not be specified so there are no cent and colname arguments and row names may not be specified so there is no rowname argument status CPXaddlazyconstraints env lp cnt nzcnt rhs sense beg ind val ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 39 CPXaddusercuts CPXaddusercuts Usage Description Return Value Synopsis Arguments Example See Also Mixed Inte
165. s that must have them like semi continuous variables CPXbasicpresolve does not create a presolved problem The interface allows the user to manually free the memory associated with the presolved problem using routine CPXfreepresolve The next optimization call or call to CPXpresolve recreates the presolved problem Modifying a Problem This section briefly discusses the mechanics of modifying a problem after presolve has been performed This discussion applies only to linear programs and mixed integer programs it does not apply to quadratic programs As noted earlier the user must indicate through the CPX_PARAM_REDUCE parameter the types of modifications that are going to be performed to the problem Recall that if primal reductions are turned off the user can add variables change the right hand side vector or loosen variable bounds without losing the presolved problem These changes are made through the standard problem modification interface CPXaddcols CPXchgrhs and CPXchgbds Recall that if dual reductions are turned off the user can add constraints to the problem change the objective function or tighten variable bounds Variable bounds are tightened through the standard interface CPXchgbds The addition of constraints or changes to the objective value must be done through the two interface routines CPXpreaddrows and CPXprechgob3 We should note that the constraints added by CPXpreaddrow
166. solution The second solution should ideally start from the basis of the first since starting from an advanced basis is usually significantly faster if the problem is only modified slightly 18 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL RESTRICTING PRESOLVE REDUCTIONS Unfortunately this scenario presents several difficulties First presolve must translate the new constraints on the original problem into constraints on the presolved problem Presolve in releases prior to 7 0 could not do this In addition the new constraints may invalidate earlier presolve reductions thus rendering the presolved problem useless for the reoptimization an example is shown in Restricting Presolve Reductions Presolve in releases prior to 7 0 had no way of disabling such reductions In the prior releases a user could either restart the optimization on the original unpresolved problem or perform a new presolve on the modified problem In the former case the reoptimization does not get the problem size reduction benefits of presolve In the latter the second optimization does not obtain the benefit of having an advanced starting solution T i gt O o gt o 2 A 0S91d paoueApY The advanced presolve interface can potentially make this and many other sequences of operations more efficient It provides facilities to restrict the set of presolve reductions performed so that subsequent problem modifications can be accommodated It also provi
167. solution has been found but before this solution replaces the incumbent This callback can be used to discard solutions that do not meet criteria beyond that of the mixed integer programming formulation For documentation of callback arguments see the routine CPXsetincumbentcallbackfunc This routine does not return a result void CPXgetincumbentcallbackfunc CPXCENVptr env int CPXPUBLIC incumbentcallback_p CPXCENVptr xenv void cbdata int wherefrom void cbhandle double objval double x int isfeas_p int useraction_p void cbhandle_p CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXPUBLIC incumbentcallback_p The address of the pointer to the current user written incumbent callback If no callback has been set the pointer evaluates to NULL void cbhandle_p The address of a variable to hold the user s private pointer CPXgetincumbentcallbackfunc env amp current_incumbentcallback amp current_handle CPXsetincumbentcallbackfunc Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 123 CPXgetkappa CPXgetkappa Usage Advanced Description The routine CPXgetkappa computes and returns an estimate of the condition number kappa Return Value If successful this routine returns an estimate of the condition number kappa Synopsis double CPXgetkappa CPXCENVptr env CPXCLPptr l
168. solve which manually invokes presolve on the supplied problem Once this routine is called on a problem the problem has a presolved problem associated with it Subsequent calls to optimization routines CPXprimopt CPXdualopt CPXbaropt CPXmipopt will use this presolved problem without repeating the presolve provided no operation that discards the presolved problem is performed in the interim The presolved problem is automatically discarded if a problem modification is performed that is incompatible with the setting Of CPX_PARAM_REDUCE further information is provided in Modifying a Problem By using the CPX_PARAM_REDUCE to restrict the types of presolve reductions CPLEX can make use of the optimal basis to the presolved problem If you set CPX_PARAM_REDUCE to restrict presolve reductions then make problem modifications that don t invalidate those reductions CPLEX will automatically use the optimal basis to the presolved problem On the other hand if the nature of the problem modifications is such that you cannot set CPX_PARAM_REDUCE you can still perform an advanced start by making the modifications calling CPXpresolve to create the new presolved problem then calling CPXcopystart passing the original model and any combination of primal and dual solutions CPLEX will crush the solutions and use them to construct a starting basis for the presolved model We should
169. space and the number of entries in indices A non negative value of surplus_p indicates that the length of the arrays was sufficient A negative value indicates that the length was insufficient and ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 131 CPXgetprotected that the routine could not complete its task In that case the routine CPXgetprotected returns the value CPXERR_NEGATIVE_SURPLUS and the value of surplus_p indicates the amount of insufficient space in the arrays Note If the value of pspace is 0 the negative of the value of surplus_p returned indicates the length needed for array indices Example status CPXgetprotected env lp amp protectcnt protectind 10 amp surplus 132 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetray CPXgetray Usage Description Return Value Synopsis Arguments Example Advanced The routine CPXget ray 1s used to find an unbounded direction or ray for a linear program where either the CPLEX primal simplex algorithm concludes that the LP is unbounded solution status CPX_UNBOUNDED An error is returned CPXERR_NOT_UNBOUNDED if this case does not hold The routine returns a zero on success and a nonzero if an error occurs int CPXgetray CPXCENVptr env CPXCLPptr lp double z CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A poin
170. sponding importModel method documented in the ILOG CPLEX Reference Manual 184 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL void presolve IloCplex Algorithm alg This member function performs Presolve on the model The enumeration alg tells Presolve which algorithm is intended to be used on the reduced model NoA1g should be specified for MIP models ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 185 lloCplex LazyConstraintCallbackl Category Nested Class Inheritance Path TloCplex CallbackI TloCplex MIPCallbackI TloCplex ControlCallbackI TloCplex CutCallbackI IloCplex LazyConstraintCallbackI Description An instance of the class IloCplex LazyConstraintCallbackI represents a user written callback in an application that uses an instance of IloCplex to solve a MIP while generating lazy constraints IloCplex calls the user written callback after solving each node LP exactly like I10Cplex CutCallback1 In fact this callback is exactly equivalent to IloCplex CutCallbackI but offers a name more consistently pointing out the difference between lazy constraints and user cuts Include File lt ilcplex ilocplex h gt Definition File lt ilcplex ilocplexi h gt Synopsis class LazyConstraintCallbackI public CutCallbackI protected LazyConstraintCallbackI y 186 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL lloCplex UserCutCallbackl Category Nested Class Inheritance
171. state double vall double val2 double 11 double ul double val3 The function write_entry is called once for each row and column in the problem Table 12 summarizes its arguments Table 12 Arguments of the write_entry Function in the Routine CPXsolwrite Type Name Meaning void info the info parameter of CPXsolwrite int aflag a for alternate optimum int num sequence number cumulative over rows and columns char name name of row or column char state state of row or column one of UL LL BS EQ FR double vall for rows row activity for columns column solution value double val2 for rows slack activitiy for columns objective coefficient double 11 for rows lower limit for columns lower bound double ul for rows upper limit for columns upper bound double val3 for rows dual value for columns reduced cost ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 171 CPXsolwrite void CPXPUBLIC sectionend void info The function sect ionend is used at the end of each header row and column section The only argument to this function is the info parameter void info A generic pointer that passes information to each of the functions called by CPXsolwrite 172 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXstrongbranch CPXstrongbranch Usage Description Return Value Synopsis Arguments Advanced The routine CPXstrongbranch co
172. steepest edge norms will be computed on the next call to CPXprimopt Synopsis void CPXkillpnorms CPXLPptr lp Arguments CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 137 CPXmdleave CPXmdleave Usage Advanced Description The routine CPXmdleave assumes that there is a resident optimal simplex basis and a resident LU factorization associated with this basis It takes as input a list of basic variables as specified by goodlist and goodlen and returns values commonly known as Driebeek penalties in the two arrays downratio andupratio For a detailed description of the conditions imposed by this function on goodlist and goodlen and the detailed meaning of the entries in downratio and upratio see the discussion in the Arguments section below Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXmdleave CPXCENVptr env CPXLPptr lp const int goodlist int goodlen double downratio double upratio Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLI routines sal pS CPXLPptr lp A pointer to a CPLEX LP problem object as returned byCPXcreateprob documented in the CPLEX Reference Manual const int goodlist An array of integers that must be
173. t the current node during MIP optimization from within a user written callback The values are from the original problem if CPX_PARAM_MIPCBREDLP is set to CPX_OFF otherwise they are from the presolved problem This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CPX_CALLBACK_MIP_HEURISTIC CPX_CALLBACK_MIP_SOLVE CPX_CALLBACK_MIP_CUT or The routine returns a zero on success and a nonzero if an error occurs int CPXgetcallbacknodeub CPXCENVptr env void cbdata int wherefrom double ub int begin int end CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines void cbdata The pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating from where the user written callback was called The parameter must be the value of wherefrom passed to the user written callback double ub An array to receive the values of the upper bound values This array must be of length at least end begin 1 If successful ub 0 through ub end begin contain the upper bound values for the current subproblem int begin An integer indicating the beginning of the range of upper bound values to be returned ILOG CPLEX 8
174. t x can be fixed to 0 it is important to note that dual information redundancy of the first dual constraint is used to formally prove it Now consider the addition of a new constraint x2 Sx Primal Dual max xl x2 x3 min 6y 5y2 st xl x2 2x3 lt 6 st yl Sy3 2 I x2 x3 lt 5 yl y y 2 1 5x1 x2 lt 0 2yl y2 gt 1 xl x2 x 2 0 yl y2 ys 2 0 Our goal is to add the appropriate constraint to the presolved problem and reoptimize Note however that the dual information presolve used to fix x to O was changed by the addition of the new constraint The first constraint in the dual is no longer guaranteed to be redundant so the original fixing is no longer valid the optimal solution is now x x2 5 x3 0 As a result CPLEX is unable to use the presolved problem to reoptimize We classify presolve reductions into several classes those that rely on primal information those that rely on dual information and those that rely on both Future addition of new constraints modifications to objective coefficients and tightening of variable bounds a special class of adding new constraints require the user to turn off dual reductions Introduction of new columns modifications to right hand side values and relaxation of 20 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL RESTRICTING PRESOLVE REDUCTIONS variable bounds a special case of modifying right hand side values require the user to turn off primal reductions
175. ter to the CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual double z The array where the unbounded direction is returned This array must be at least as large as the number of columns in the problem object As an illustration for a linear program of the form Minimize cx Subject to Ax b x 20 If the CPLEX primal simplex algorithm completes optimization with a solution status of CPX_UNBOUNDED the vector z returned by CPXget ray would satisfy cz lt 0 Az 0 z20 if computations could be carried out in exact arithmetic status CPXgetray env lp z ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 133 CPXgetredIp CPXgetredlp Usage Advanced Description The routine CPXgetredlp returns a pointer for the presolved problem It returns NULL if the problem is not presolved or if all the columns and rows are removed by presolve Generally the returned pointer may be used only in CPLEX Callable Library query routines such as CPXsolution Or CPXgetrows The presolved problem must not be modified Any modifications must be done on the original problem If CPX_PARAM_REDUCE is set appropriately the modifications are automatically carried out on the presolved problem at the same time Optimization and query routines can be used on the presolved problem Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXget
176. the new value of the bound int rent An integer that indicates the number of constraints for the branch int nzcnt An integer that indicates the number of nonzero constraint coefficients for the branch This specifies the length of the arrays rmatind and rmatval const double rhs An array of length rent containing the right hand side term for each constraint for the branch const char sense An array of length rent containing the sense of each constraint to be added for the branch sense i L Sconstraint sense i E constraint sense i G 2 constraint ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 51 CPXbranchcallbackbranchgeneral const int rmatbeg const int rmatind const double rmatval Arrays that describe the constraints for the branch The format is similar to the format used to describe the constraint matrix in the routine CPXaddrows Every row must be stored in sequential locations in this array from position rmatbeg i to rmatbeg i 1 1 or from rmatbeg i tonzcnt 1if i rcnt 1 Each entry rmatind i indicates the column index of the corresponding coefficient rmatval 1 All rows must be contiguous and rmatbeg 0 must be 0 void userhandle A pointer to user private data that should be associated with the node created by this branch May be NULL int seqnum_p A pointer to an integer that on return will contain the sequence number that CPLEX has assigned to the node creat
177. these advanced presolve routines The second option is to set CPX_PARAM_MIPCBREDLP to CPX_OFF 0 thus requesting that the callback routines work exclusively with original variables CPLEX automatically translates the data between original and presolved data While the second option is simpler the first provides more control These two options will be revisited at several points in this chapter Heuristic Callback The first user callback we consider is the heuristic callback The first step in using this callback is to call CPXsetheuristiccallbackfunc with a pointer to a callback function and optionally a pointer to user private data as arguments We refer the reader to advanced example admipex2 c for further details of how this callback is used Once this routine has been called CPLEX calls the user callback function at every viable node in the branch amp cut tree we call a node viable if its LP relaxation is feasible and its relaxation objective value is better than that of the best available integer solution The user callback routine is called with the solution vector for the current relaxation as input The callback function should return a feasible solution vector if one is found as output Q o gt 2e y 23 0 3 0 za z O U The advanced MIP control interface provides several routines that allow the user callback to gather information that may be useful in finding heuristic solutions The routines CPXgetcall
178. tion 2 CPX_CALLBACK_SET Proceed with optimization ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 163 CPXsetincumbentcallbackfunc Example status CPXsetincumbentcallbackfunc env myincumbentcheck mydata See Also CPXgetincumbentcallbackfunc Advanced MIP Control Interface 164 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXsetnodecallbackfunc CPXsetnodecallbackfunc Usage Mixed Integer Users Only Description The routine CPXsetnodecallbackfunc sets and modifies the user written callback to be called during MIP optimization after ILOG CPLEX has selected a node to explore but before this exploration is carried out The callback routine can change the node selected by ILOG CPLEX to a node selected by the user Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXsetnodecallbackfune CPXENVptr env int CPXPUBLIC nodecallback CPXCENVptr env void cbdata int wherefrom void cbhandle int nodeindex_p int useraction_p void cbhandle Arguments CPXENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines int CPXPUBLIC nodecallback A pointer to the current user written node callback If no callback has been set the pointer evaluates to NULL void cbhandle A pointer to user private data This pointer is passed to the user written node callback Callback Description ILOG
179. to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines void cbdata The cbdata pointer passed to the user written callback This parameter must be the value of cbdata passed to the user written callback int wherefrom An integer value indicating where the user written callback was called from This parameter must be the value of the wherefrom passed to the user written callback 100 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXgetcallbacknodelp CPXLPptr curlp_p The 1p pointer indicating the current subproblem If no subproblem is defined the pointer is set to NULL Example status CPXgetcallbacknodelp env cbdata amp curlp See Also The examples adqmipex1 c and admipex6 c CPXsetbranchcallbackfunc CPXsetcutcallbackfunc CPXsetheuristiccallbackfunc CPXsetsolvecallbackfunc Advanced MIP Control Interface ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 101 CPXgetcallbacknodeobjval CPXgetcallbacknodeobjval Usage Description Return Value Synopsis Arguments Example See Also 102 Mixed Integer Users Only The routine CPXgetcallbacknodeobjval is used to get the objective value for the subproblem at the current node during MIP optimization from within a user written callback This routine may be called only when the value of the wherefrom argument is one of CPX_CALLBACK_MIP CPX_CALLBACK_MIP_BRANCH CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_NODE CP
180. tr env CPXLPptr lp const int goodlist int goodlen double downpen double uppen int itlim CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLI routines E x lt CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 173 CPXstrongbranch 174 const int goodlist An array of integers The length of the array must be at least goodlen As in other ILOG CPLEX Callable Library routines row variables in goodlist are specified by the negative of row index shifted down by one that is rowindex 1 int goodlen An integer indicating the number of entries in goodlist double downpen An array containing values that are the result of the downward fix of branching variables in dual steepest edge iterations carried out by CPXstrongbranch The length of the array must be at least goodlen double uppen An array containing values that are the result of the upward fix of branching variables in dual steepest edge iterations carried out by CPXstrongbranch The length of the array must be at least goodlen int itlim An integer indicating the limit on the number of dual steepest edge iterations carried out by CPXstrongbranch on each LP ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXtightenbds CPXtightenbds Usage Descripti
181. trix has the internal index 3 1 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 175 CPXtightenbds const char lu An array containing characters indicating whether the corresponding entry in the array ba specifies the lower or upper bound on column indices j The allocated length of the array is cnt Table 13 summarizes the values that entries in this array may assume Table 13 Bound Indicators in the Argument 1u of CPXtightenbds Value of lu j Meaning for bd j U bd j is an upper bound L bd 3 is a lower bound B bd 3 is the lower and upper bound const double bd An array containing the new values of the upper or lower bounds of the variables present in the array indices The allocated length of the array is cnt Example status CPXtightenbds env lp cnt indices lu bd 176 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXuncrushform CPXuncrushform Usage Advanced Description The routine CPXuncrushform uncrushes a linear formula of the presolved problem to Return Value Synopsis Arguments a linear formula of the original problem The routine returns a zero on success and a nonzero if an error occurs int CPXuncrushform CPXCENVptr env CPXCLPptr lp int plen const int pind const double pval int len_p double offset_p int ind double val CPXCENVptr env sal pS The pointer to the CPLEX environment as returned by one of the CPXopenC
182. ualopt 22 31 58 73 79 136 139 CPXERR_NEGATIVE_SURPLUS 131 132 CPXERR_NO_SPACE 67 CPXERR_NODE_ON_DISK 93 112 CPXERR_NOT_UNBOUNDED 133 CPXERR_PRESLV_INF 21 CPXERR_PRESLV_UNBD 21 CPXfreelazyconstraints 75 CPXfreepresolve 21 76 CPXfreeprob 21 CPXfreeusercuts 40 77 CPXftran 78 CPXgetbasednorms 79 CPXgetbhead 81 CPXgetbranchcallbackfunc 82 CPXgetcallbackctype 84 CPXgetcallbackglobal lb 27 28 86 CPXgetcallbackglobalub 27 88 CPXgetcallbackincumbent 27 90 CPXgetcallbacklp 27 28 92 CPXgetcallbacknodeinfo 93 112 166 CPXgetcallbacknodeintfeas 29 30 96 154 CPXgetcallbacknodelb 27 102 CPXgetcallbacknodelp 27 92 100 154 167 CPXgetcallbacknodeobjval 102 202 ILOG CPLEX 8 1 000 001 00400 0 00010000 00000000000 CO OO 10000000 OA PXgetcallbacknodestat 103 PXgetcallbacknodeub 27 104 PXgetcallbacknodex 28 106 PXgetcallbackorder 30 108 PXgetcallbackpseudocosts 29 110 PXgetcallbacksosinfo 114 154 PXgetcutcallbackfunc 116 PXgetdeletenodecallbackfunc 117 PXgetdnorms 118 PXgetExactkappa 119 PXgetheuristiccallbackfunc 120 PXgetijdiv 121 PXgetijrow 122 PXgetincumbentcallbackfunc 123 PXgetkappa 124 PXgetnodecallbackfunc 125 PXgetnumcols 79 138 PXgetnumrows 79 PXgetobjoffset 126 PXgetpi 179 PXgetpnorms 127 PXgetprestat 128 PXgetprotected 131 132 PXgetray 133 PXgetredlp 23 134 PXgetrows 92 100 134 PXgetsolvecallbackfunc 135 PXgetstat 73 PXgetx 66 18
183. uld be specified by CPXaddusercuts but in 7 1 CPXaddusercuts should be used only to specify cuts that is additional constraints that are implied by the constraints in the MIP formulation but that may be helpful in obtaining a proved optimal integer solution CPXaddlazyconstraints and CPXfreelazyconstraints have the same arguments as the user cut functions 194 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPLEX 7 1 ADVANCED FEATURES RELEASE NOTES In the Concert Technology Library the following are added to provide capabilities for lazy constraints IloCplex addLazyConstraint IloCplex addLazyConstraints IloCplex clearLazyConstraints IloCplex LazyConstraintCallbackI ILOLAZYCONSTRAINTCALLBACK macro CPXmipopt will return with an error when the advanced presolve parameters are set in conflict with the presence of user cuts and or lazy constraints For lazy constraints CPX_PARAM_REDUCE must be set to CPX_PREREDUCE_NOPRIMALORDUAL or CPX_PREREDUCE_PRIMALONLY For user cuts CPX_PARAM_REDUCE must be set to CPX_PREREDUCE_NOPRIMALORDUAL Or CPX_PREREDUCE_PRIMALONLY or CPX_PARAM_PRELINEAR must be set to 1 A new chapter has been added to provide information on using the user cuts and lazy restraints See Chapter 3 User Cut and Lazy Constraint Pools ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 195 CPLEX 6 6 TO 7 0 ADVANCED
184. urrent node It possibly also gathers other information about the problem through CPXgetcallbacklp CPXgetcallbackgloballb etc It then calls a user separation routine to identify violated user cuts These cuts are then added to the problem by calling CPXcutcallbackada and the callback returns Local cuts that is cuts that apply to the subtree of which the current node is the root can be added by calling CPXcutcallbackaddlocal At this point it is important to draw a distinction between the two different types of constraints that can be added through the cut callback interface The first type is the traditional MIP cutting plane which is a constraint that can be derived from other constraints in the problem and does not cut off any integer feasible solutions The second is a lazy constraint which is a constraint that can not be derived from other constraints and potentially cuts off integer feasible solutions Either type of constraint can be added through the cut callback interface As with the heuristic callback the user can choose whether to work with presolved values or original values If the user chooses to work with original values a few parameters must be modified If the user adds only cutting planes on the original problem the user must set advanced presolve parameter CPX_PARAM_PRELINEAR to CPX_OFF 0 This parameter ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL BRANCH SELECTION CALLBACK forbi
185. user is examined for logical reduction opportunities The goal is to reduce the size of the problem passed to the requested optimizer A reduction in problem size typically translates to a reduction in total run time even including the time spent in presolve itself ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 17 INTRODUCTION TO PRESOLVE Consider scorpion mps from NETLIB as an example CPLEX gt disp pr st Problem name scorpion mps Constraints 388 Less 48 Greater 60 Equal 280 Variables 358 Constraint nonzeros 1426 Objective nonzeros 282 RHS nonzeros 76 CPLEX gt optimize Tried aggregator 1 time LP Presolve eliminated 138 rows and 82 columns Aggregator did 193 substitutions Reduced LP has 57 rows 83 columns and 327 nonzeros Presolve time 0 00 sec Iteration log Iteration 1 Dual objective 317 965093 Dual Optimal Objective 1 8781248227e 03 Solution time 0 01 sec Iterations 54 0 CPLEX is presented with a problem with 388 constraints and 358 variables and after presolve the dual simplex method is presented with a problem with 57 constraints and 83 variables Dual simplex solves this problem and passes the solution back to the presolve routine which then unpresolves the solution to produce a solution to the original problem During this process presolve builds an entirely new presolved problem and stores enough information to translate a solution to this problem back to a solution t
186. user private data int useraction_p A pointer to an integer indicating the to be taken on completion of the user callback Table 11 summarizes possible actions Table 11 Actions to be Taken after a User Written Solve Callback Value Symbolic Constant Action 0 CPX_CALLBACK_DEFAULT Use ILOG CPLEX subproblem optimizer 1 CPX_CALLBACK_FAIL Exit optimization 2 CPX_CALLBACK_SET The subproblem has been solved in the callback Example status CPXsetsolvecallbackfunc env mysolvefunc mydata See Also The example admipexl c CPXgetsolvecallbackfunc CPXgetcallbacknodelp Advanced MIP Control Interface 168 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL CPXslackfromx CPXslackfromx Usage Advanced Description The routine CPXslackfromx computes an array of slack values from primal solution values Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXslackfromx CPXCENVptr env CPXCLPptr lp const double x double slack Arguments CPXCENVptr env The pointer to the CPLEX environment as returned by one of the CPXopenCPLEX routines CPXCLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual const double x An array that contains primal solution x values for the problem as returned by routines such as CPXcrushx and CPXuncrushx The array must be of l
187. ut process 26 building 18 changing objective function coefficients 145 freeing 21 freeing from LP problem object 76 freeing memory 24 obtaining status information 128 retaining 21 uncrushing dual solution 149 179 uncrushing solution 180 primal solution values in subproblem 160 primal steepest edge norms 61 127 137 ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 205 INDEX primal variable values accessing for subproblem from user written callback 106 priority order accessing from user written callback 108 pseudo cost values accessing from user written callback 110 R reduced cost computing from dual values 72 right hand side RHS cuts and 67 70 vector 53 78 row adding to problem object and presolved problem 144 detecting redundant 41 S scaling 54 55 181 sequence number 94 154 166 accessing 94 166 cuts and 154 setting incumbent callback 162 slack 81 computing from primal solution values 169 in basis header 81 pivoting into basis 141 primal steepest edge norms 127 solution callbacks and 120 159 crushing from original problem to presolved 66 discarding 123 162 file 170 heuristics and 120 159 integer 120 159 uncrushing from presolved problem to original 180 using advanced presolved 22 vector 78 writing 170 solve callback 31 special ordered set SOS 94 152 accessing callback information 114 indexing 114 member of 114 numbering conventions 114 priority 115 steepest edge iterations 1
188. ws When a user cut pool is present the CPX_PARAM_PRELINEAR parameter must be set to 0 When a lazy constraint pool is present the CPX_PARAM_REDUCE parameter must be set to CPX_PREREDUCE_PRIMALONLY 1 or CPX_PREREDUCE_NOPRIMALORDUAL 0 If these parameters do not have these values the error CPXERR_PRESOLVE_BAD_PARAM will be issued when CPXmipopt is called The Concert Technology CPLEX Library will automatically handle these parameter settings ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 35 DELETING USER CUTS AND LAZY CONSTRAINTS Deleting User Cuts and Lazy Constraints 36 The user cut and lazy constraint pools are cleared by calling the routines CPXfreeusercuts and CPXfreelazyconstraints Clearing the pools will not change the MIP solution The Concert Technology routines are IloCplex clearUserCuts and IloCplex clearLazyConstraints ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL Part Il Routines CPXaddlazyconstraints CPXaddlazyconstraints Usage Description Return Value Synopsis Arguments Example Mixed Integer Users Only The routine CPXaddlazyconstraints is used to add constraints to the list of constraints that should be added to the LP subproblem of a MIP optimization if they are violated CPLEX handles addition of the constraints and makes sure that all integer solutions satisfy all the constraint
189. y constraints added to the invoking IloCplex object with the methods addLazyConstraint and addLazyConstraints void clearUserCuts This member function deletes all user cuts added to the invoking IloCplex object with the methods addUserCut and addUserCuts void freePresolve This member function frees the presolved problem Under the default setting of parameter Reduce the presolved problem is freed when an optimal solution is found however it is not freed if Reduce has been set to 1 primal reductions or to 2 dual reductions In these instances the function freePresolve can be used when necessary to free it manually ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 183 IloExtractable getDiverging This member function returns the diverging variable or constraint in a case where the primal Simplex algorithm has determined the problem to be infeasible The returned extractable is either an I1oNumVar or an IloRange object extracted to the invoking IloCplex optimizer it is of type IloNumVar if the diverging column corresponds to a variable or of type IloRange if the diverging column corresponds to the slack variable of a constraint void getRay This member function returns an unbounded direction corresponding to the present basis for an LP that has been determined to be an unbounded problem void importModel IloModel amp m const char filename TloObjective amp obj TloNumVarArray vars TloRangeA
190. y one of CPX_CALLBACK_MIP_CUT CPX_CALLBACK_MIP_INCUMBENT CPX_CALLBACK_MIP_HEURISTIC or CPX_CALLBACK_MIP_BRANCH only the current node can be queried This is done by specifying a nodeindex value of 0 Other values of the wherefrom argument are invalid for this routine an invalid nodeindex value or wherefrom argument value will result in an error return value Notes The values returned for CPX_CALLBACK_INFO_NODE_SIINF and CPX_CALLBACK_INFO_NODE_NIINF for the current node are the values that applied to the node when it was stored and thus before the branch was solved As a result these values should not be used to assess the feasibility of the node This routine cannot retrieve information about nodes that have been moved to node files For more information about node files see the ILOG CPLEX User s Manual If the argument nodeindex refers to a node in a node file CPXgetnodecallbackinfo returns the value CPXERR_NODE_ON_DISK Nodes still in memory have the lowest index numbers so a user can loop through the nodes until CPXgetcallbacknodeinfo returns an error and then exit the loop Return Value The routine returns a zero on success and a nonzero if an error occurs The return value CPXERR_NODE_ON_DISK indicates an attempt to access a node currently located in a node file on disk Synopsis int CPXgetcallbacknodeinfo CPXCENVptr env void cbdata int wherefrom int nodeindex int
191. y routine However any starting point information for the current solution such as an associated basis is retained Return Value The routine returns a zero on success and a nonzero if an error occurs Synopsis int CPXunscaleprob CPXCENVptr env CPXLPptr lp Arguments CPXCENVptr env The pointer to the ILOG CPLEX environment as returned by one of the CPXopenCPLEX routines CPXLPptr lp A pointer to a CPLEX LP problem object as returned by CPXcreateprob documented in the CPLEX Reference Manual ILOG CPLEX 8 1 ADVANCED REFERENCE MANUAL 181 lloCplex Category Inheritance Path Description Member Functions Handle Class IloAlgorithm IloCplex We include here additional methods of 110Cp1ex For a description of the IloCplex class see the ILOG CPLEX Reference Manual IloRange addLazyConstraint IloRange rng This member function adds rng as a lazy constraint to the invoking IloCplex object The range rng is copied into the lazy constraint pool the rng itself is not part of the pool so changes to rng after it has been copied into the lazy constraint pool will not affect the lazy constraint pool Lazy constraints added with addLazyConstraint are typically constraints of the model that are not expected to be violated when left out The idea behind this is that the LPs that are solved when solving the MIP can be kept smaller when these constraints are not included IloCplex will however include

Download Pdf Manuals

Related Search

Related Contents

TCP/IP Ethernet Communications for the Series 90 PLC User`s Manual  取扱説明書 - 日東工業株式会社 N-TEC  Lindy 20988 User's Manual  INVERTEC V270-T AC/DC  ANNEXE 8 - Prévoyance  "取扱説明書"  〈取扱説明書〉 A B C - 株式会社リッチェル ペット用品  MSI WindPad Enjoy 10 Plus Operating Manual  listado de formatos, instructivos y anexos  Toshiba Satellite C55t-A5218  

Copyright © All rights reserved. Failed to retrieve file