This paper has been mechanically scanned. Some errors may have
Contents
1. In addition the user is allowed to change the period length and hence change the entire arrival volume distribution in a single step Suppose the user intends to change period length from the current 42 time slices to 12 time slices and redefine the distribution of the traffic volume as follows Time Slice Volume 300 300 300 300 400 400 400 500 600 10 700 11 700 12 700 o 00 rd GO LI PWD To do so the following steps should be followed after entering the menu Traffic Volume Distribution shown in Fig 3 4 1 Enter a the screen will prompt Data range must be from 10 to 45 Enter total time period gt 2 Enter 12 the screen will prompt Enter traffic volume at tim slice 1 OR press ENTER for default value default 100 00 gt 3 Enter 300 Repeat Step 3 until done If the default value shows the volume as wanted press ENTER otherwise enter a value After time slice 1 in which the default value is 100 the default value always duplicates the value in the previous time slice For consecutive identical arrival volumes the user needs to enter the value only once and simply press ENTER to duplicate 11 thereafter become A RR PR TO 1 time slice 5 minute Tim li Volumes co e oO db WwW Pa 300 300 300 300 400 400 400 500 9 600 10 700 11 700 12 700 00 00 00 00 00 00 00 00 00 00 00 002. AWTRT1 Average system wait Route 1 iterations 2 AWTRT2 Average system wait Route 2 iterations 2 AWTRT3 Average system wait Route 3 iterations 2 AWTTOT Average system wait total iterations 2 4 PWTRT1 Percentile system wait Route 1 iterations 2 4 6 8 PWIRT2 Percentile system wait Route 2 iterations 2 4 6 8 PWIRT3 Percentile system wait Route 3 iterations 2 4 6 8 PWITOT Percentile system wait total iterations 2 4 6 8 1 RWTRT1 Recurrent system wait Route 1 iterations 2 4 RWIRT2 Recurrent system wait Route 2 iterations 2 4 2 4 4 6 RWTRT3 Recurrent system wait Route 3 iterations RWITOT Recurrent system wait total iterations 2 4 6 RWTUPS Recurrent system wait upstream iterations 2 4 6 8 10 Figures 5 1 5 5 show five preset graphs CUMTOT AWTTOT RWITOT PWITOT and RWTUPS for the example simulation Notice that the graphs show how the gueues evolve over time both from iteration to iteration and time slice to time slice By comparing AWTTOT to RWTTOT the extent to which incidents affect delay can also be assessed For instance note that incidents have a much larger impact at the end of the rush hour than at the start In addition to the preset graphs users may choose to create individualized graphs These can be saved in the Lotshell wk1 file for reuse on different data sets However to reduce memory reguired for future simulations do not save your simulation output under the file name
3. VOLUM Tim After the change the screen in Fig 3 4 will DISTRIBUTION E interval li Volum Time Slice Volumes 0 save and exit to Edit Input Data File menu Enter a number and press ENTER to change a single slice 0 12 or enter a for change all entries and or number of slices gt The following is a full description of the data in each category including the ranges dimensions units and the representations A Route Characteristics O Probability that no incidents occur Range 0 00 1 00 Dimen sion scalar Unit none Probability that no new incident occurs in any time slice O Free flow travel time Range 0 00 99999 00 Dimension 1 3 Unit minute s Time reguired to travel through route i with no wait in gueue O Bottleneck capacity Range 0 00 99999 00 Dimension 1 3 Unit vehicles hour Nominal capacity by route in the absence of incidents O Ups Range tream capacity 0 00 99999 00 Dimension scalar Unit vehicles hour 12 Total upstream capacity among all routes Program assumes no incidents occur upstream The upstream capacity must exceed the combined capacity of the bottleneck routes O Toll cost Range 0 00 99999 00 Dimension 1 3 1 3 Units dollar te This toll is route specific but does not vary over time It is used in the cost equation for Type 3 travelers cost minimizers Therefore it primarily af
4. the duration and magnitude are simulated and the bottleneck capacity is decremented as appropriate However if one incident occurs soon after another the incident with the largest magnitude defines the capacity until it is cleared up To illustrate these concepts consider the following example pertaining to a 3 lane freeway with nominal capacity of 6000 vehicles hour Time Slice Incident Duration Capacity Loss Current Capacity slices veh hr 1 No as EO 6000 2 No tis Za 6000 3 Yes 3 33 4000 4 No Ee Ze 4000 5 Yes 4 20 4000 6 No KO E 4800 7 Yes 1 50 3000 8 No BE ZO 4800 9 No a lt 6000 10 No ae 6000 As the example demonstrates the current capacity is defined by the outstanding incident with the biggest capacity loss In time slice 5 the first incident has the biggest loss and the capacity is 4000 hour However the first incident is cleared at the end of slice 5 so the capacity in time slice 6 is 4800 as defined by the second incident The queue size at the end of a time slice is calculated in the following manner Let A number of arrivals in slice i C capacity in slice i vehicles hour Qi queue size at the end of slice i n number of time slices per hour N number of time slices in simulation Then Qi max Aj C n Q 4 0 The first part of the equation applies when the bottleneck operates at capacity and queues exist while the second applies when the bott
5. traveler types If desired all travelers can be allocated to the first type in which case the distribution of arr ivals among the time slices remains constant from iteration to iteration Within each type different travelers have different des ired arrival or departure times as specified by the user CHAPTER 3 CREATING THE INPUT DATA The input data for BTS can be divided into five categories route characteristics system breakdown driver choice behavior traffic volume distribution and the simulation control data The data from the first four categories can be modified through the selection Edit Input Data File in the main menu of BTS The last category can be modified before simulation by choosing the selection Run Simulation Program Initially the user is provided with a set of data shown in Fig 3 1 3 4 With the selection Edit Input Data File the user can change any data if so desired The same basic procedure is used for making all changes When finished with a screen enter 0 and all changes will be saved For example after entering the main menu of BTS to change the capacity for Route 1 from 3600 to 4700 veh hr the user needs to go through the following steps 1 Enter 1 to choose the selection Edit Input Data File 2 Enter 1 to choose the selection Route Characteristics The panel in Fig 3 1 will appear on the s
6. Lotshell wkl 20 Cumulative Vehicles Thousands Waiting Time 11 ou o Y 00 0 o A N UQ o L i D N 6 it 16 21 26 31 36 41 Time Slice Fig 5 1 Cumulative Arrivals Total Time Slice Fig 5 2 Average Waiting Time Total 21 Waiting Time Waiting Time minutes 1 6 11 16 21 25 31 36 41 Time Slice Fig 5 3 Recurrent Waiting Time Total gt a a op ou o vu o o o N GO d oon ON au PU DN Oo Time Slice Fig 5 4 Percentile Waiting Time Total 22 Waiting Time minutes Note b gt Oek O Iu I Fig in Fig Iteration Iteration Iteration Iteration Iteration oF 5 Ge bk GO GO E bd Time Slice Recurrent Waiting Time Ge 23 Upstream CHAPTER 6 CONCLUSION BTS can be used to evaluate a variety of changes in highway design to improve bottlenecks such as 1 addition of highway lanes 2 addition of automated or HOV lanes or 3 incident management strategies to reduce the frequency duration and magnitude of incidents The current version of BTS is experimental In a future issue BTS will be enhanced to include incident dependencies weather conditions and randomly varying traffic volumes 24
7. behavior including changes in departure time and changes in route Interfaces with Lotus 1 2 3 to provide graphs of cumulative arrivals and departure and to demonstrate queueing over time BTS operates on an IBM PC or compatibles with two EXE files three COM files and one DAT file A fully operative BTS including output files requires about 250K of work disk space To obtain a copy of BIS contact Randolph Hall at the Department of Industrial Engineering amp Operations Research University of California at Berkeley 94720 CHAPTER 2 THEORY BEHIND BTS BTS was developed to illustrate how a freeway bottleneck behaves over time as traffic levels and freeway performance change By focusing on a single bottleneck rather than an entire freeway corridor or network BTS provides more details about roadway performance than previously possible In particular through randomly simulating the occurence duration and magnitude of freeway incidents BTS calculates average recurrent delay and average non recurrent delay This information is especially valuable in evaluating the merits of incident management strategies To run BTS the user must first specify an initial demand pattern as well as highway characteristics BTS uses this input to evaluate travel time performance by time slice If desired BTS can iteratively adjust the demand pattern in response to the travel t
8. 12 400 00 27 300 00 42 250 00 13 550 00 28 300 00 14 550 00 29 300 00 T5 550 00 30 300 00 0 save and exit to Edit Input Data File menu Enter a number and press ENTER to change a single slice 0 42 or enter a for change all entries and or number of slices gt Fig 3 4 Selection 4 Traffic Volume Distribution 10 ROUT tr CHAR 1 Free flow travel time minutes 2 Bottleneck capacity veh hour 5 Upstream capacity veh hour 8 Toll dollars Route 9 d 0 1 1 1 2 2 3 2 4 2 5 3 6 3 7 3 Edit Input Data 0 17 0 save and exit to Enter a number for change Fig 3 1 Selection 1 Probability that no incident occurs AC TPE Reh S5 Td CAS 0 99 Rte Rte 2 Rte 3 17 50 3 12 50 4 7 50 3600 6 2800 7 2000 8800 Type Value 0 00 2 0 00 3 2 00 0 00 2 0 00 3 2 00 0 00 2 0 00 3 2 00 File menu and press ENTER gt Route Characteristics SYSTEM BREAKDOWN Duration of breakdown uniform distribution 1 Lower bound 1 00 2 Upper bound 12 00 Capacity loss distribution Level of capacity loss Probabilitv 3 10 0 45 4 20 0 35 5 30 0 05 6 40 0 05 7 50 0 05 8 60 0 05 9 10 0 00 10 80 0 00 11 90 0 00 12 100 0 00 0 save and exit to Edit Input Data File menu Enter a number for change 0 12 and press ENTER gt Fig 3 2 Selection 2 System Breakdown 9
9. N Saas e mies aa es ea ATO i al a aa me al al da b Traffic ASSIGNMENT Ro AE OE OR OA CHAPTER E E OO EO E CHAPTER 4 AN EXAMPLE RUN CHAPTER CHAPTER 6 CONCLUSION B l Arrival Time Assignment 999m nnr B 2 Route Assignment 3 CREATING THE INPUT DATA Y Y Y YW ane YN KK Route Characteristics as ee Ye E b ra Ga E e e E b era a System Breakdown Ae Driver Chotas Behavior ese hs ier Bee WYW Ee Re e erda Traffic Volume Distribution o o oooooooooo Simultatton Control Data lt t KTE a Cee ee ee UU WT 5 EVALUATION OF BOTTLENECK PERFORMANCE WITH LOTUS 1 2 3 N HUE NO 12 13 13 14 15 16 CHAPTER 1 INTRODUCTION BTS Bottleneck Traffic Simulator is a macroscopic tool for simulating the performance of freeway bottlenecks BTS can be used to measure the travel time benefits of changes in roadwa design through 1 addition of capacity 2 increase in trave speed or 3 improvement in roadway reliability The unique features of BTS include O Allows non standard roadway characteristics which might be created through highway automation O Evaluates lane stoppages or other incidents through random simulation O Determines both recurrent and non recurrent traffic delay a Measures impact of growth in traffic level Q Accounts for changes in traveler
10. This paper has been mechanically scanned Some errors may have been inadvertently introduced Program on Advanced Technology for the Highway INSTITUTE OF TRANSPORTATION STUDIES UNIVERSITY OF CALIFORNIA AT BERKELEY BTS Version 1 O Bottleneck Traffic Simulator User s Manual Wei Hua Lin Randolph W Hall PATH Working Paper UCB ITS PWP 91 1 This work was performed as part of the Program on Advanced Technology for the Highway PATH of the University of California in cooperation with the State of California Business and Transportation Agency Department of Transportation and the United States Department of Transportation Federal Highway Administration The contents of this report reflect the views of the authors who are responsible for the facts and the accuracy of the data presented herein The contents do not necessarily reflect the official views or policies of the State of California This report does not constitute a standard specification or regulation March 1991 ISSN 10551417 BTS Version 1 0 Bottleneck Traffic Simulator User s Manual by Wei Hua Lin Randolph W Hall March 1991 Program on Advanced Technology for the Highway Institute of Transportation Studies University of California at Berkeley CHAPTER E TABLE OF CONTENTS INTRODUCTION CHAPTER 2 THEORY BEHIND BTS Sg ge Ziz nin sa A Per F Tormance Eva lu at TO
11. change routes at any iteration Now suppose that the following costs are available for Type 3 travelers Desired cost Current Assignment Time Slice Rte 1 Rte 2 Rte 3 Rte 1 Rte 2 Rte 3 auto auto 9 7 6 5 140 NE 60 NE 0 NE 160 AE 0 AE 40 AE For AE vehicles Route 3 is optimal and for NE vehicles Route 2 is optimal Therefore in the next iteration vehicles will be assigned as follows for the time slice Route 1 Not automated 140 20 140 112 20 shift to route 2 Automated 160 20 160 128 20 shift to route 3 Route 2 Not automated 60 20 140 88 20 come from route 1 Automated 0 Route 3 Not automated 0 Automated 40 20 160 72 20 come from route 1 The assignment to routes for Type 1 and 2 travelers is done in a Similar fashion Running BTS over several iterations reveals how traveler behavior volves over time However BTS does not guarantee that route and 6 arrival time choices converge toward an equilibrium Rather BTS is designed to mimic different types of behavior which may or may not converge After an iteration is completed the number of arrivals within each class for each time slice is incremented by a user specified proportion representing annual growth in highway traffic However if desired the number of arrivals can be held constant from iteration to iteration The user specifies how demand should be allocated to the different
12. creen 3 Enter 5 to choose the option for changing the capacity for Route 1 The screen will prompt Data range must be from 0 00 to 99999 00 Change from 3600 00 to gt 4 Enter 4700 When the Route Characteristics screen reappears the capacity for Route 1 is replaced by 4700 The user can change any other data shown in Fig 3 1 to 3 4 using these steps DRIVER C HOI CCE BEHAVIOR 1 Arrival multiplier for each iteration 1 05 Type 3 cost hr 2 Late 24 00 3 Early 6 00 4 Oueue 12 00 Initial routing Route Type Proportion 5 1 0 45 6 1 2 0 00 7 1 3 0 00 8 2 0 30 9 2 2 0 00 10 2 3 0 00 11 3 0 25 12 3 2 0 00 13 3 3 0 00 Proportion of travelers that would change route 14 Rte 1 0 20 15 Rte 2 0 20 16 Rte 3 0 20 17 Pet queue for Type 2 travelers 95 18 Prop with automation 0 25 0 save and exit to Edit Input Data File menu Enter a number for change 0 18 and press ENTER gt Fig 3 3 Selection 3 Driver Choice Behavior TRAEFIGC VOLUME DOES GE ZE D B EE ZE ZE Orn d 1 time slice 5 minute interval 1 200 00 16 550 00 31 300 00 2 200 00 17 550 00 32 300 00 3 200 00 18 550 00 33 300 00 4 200 00 19 550 00 34 300 00 5 300 00 20 550 00 35 300 00 6 300 00 21 400 00 36 300 00 7 300 00 22 400 00 37 250 00 8 400 00 23 400 00 38 250 00 9 400 00 24 400 00 39 250 00 10 400 00 25 400 00 40 250 00 TL 400 00 26 300 00 41 250 00
13. e and arrival time choices do not change In addition free flow travel time capacities Type 3 cost arrival multiplier for each iteration and percentile queue for Type 2 vehicles can be changed after selecting Run Simulation Program 15 CHAPTER 4 AN EXAMPLE RUN Before running BTS the files BTS EXE QM EXE INPUT ZON VIEW COM EDIT COM and BTS DAT must be on the run time disk The steps to run the simulator using the data displayed in Fig 3 1 3 4 in Chapter 3 are shown as follows I 1 Type command BTS and press ENTER The following menu displays BOTTLENECK TRAFFIC SIMULATOR Edit Input Data File Run Simulation Program Display Output File Exit to Dos GO LA b Ep Enter your choice 0 3 and press ENTER gt 2 Select choice two You will see the following panel EDIT DATA BEFORE SIMULATION 1 Probability that no incident occurs 0 99 Rte Bte Rte 3 Free flow travel time minutes 2 17 50 3 22 50 4 7 50 Bottleneck capacity veh hour 5 3600 6 2800 7 2000 8 Upstream capacity veh hour 8800 9 Arrival multiplier for each iteration 1 05 Type 3 cost dollars hour 10 Late 24 00 11 Early 6 00 12 Queue 12 00 13 Percentile queue for Type 2 travelers 95 14 Number of iterations 10 15 Number of runs in each iteration 50 0 save and start simulation Ent
14. er a number for change 0 15 and press ENTER gt 16 3 4 at this time you can change any data shown in the panel if so desired When done enter 0 All of the modified data will be saved and the simulation will start The screen will display Iteration 1 walt When the simulation is done the following panel is shown on the screen which provides a summary of collected at the end of each iteration zr Summary Statistics the statistics Free Flow System Queue Time per Veh Upstream Total Travel Time Recurrent Total Total Recurrent Iteration Arrival Per Veh Per Veh Average 95 Per Veh vehicles minutes min min min min 1 14900 13 500 0 000 0 066 0 000 0 000 2 15645 12 862 0 000 0 061 0 232 0 000 3 16427 12 352 0 558 0 822 1 808 0 000 4 17249 11 984 2 066 2207 2 865 0 000 5 18111 11 852 3 174 3 835 7 231 0 000 6 19017 12 076 2 431 3 420 8 784 0 000 7 19967 12 198 2 844 3 670 7 560 0 034 8 20966 12 207 5 406 6 492 11 851 0 396 9 22014 12 381 5 350 6 300 so 0 922 10 2 EDD 12 487 7 064 8 043 11 810 1 672 Press ENTER to exit gt This data summarizes the bottleneck performance routes among all three Press enter to bring the main menu on the screen The output files Report out and Lotus prn are now created These files provide detailed simulation output that can be used for further analysis The user can either display the output file
15. ers have fixed arrival times only the arrival times for Type 2 and 3 travelers are influenced by the data In order to arrive at work on time suppose that a Type 2 traveler must depart from the bottleneck by the end of slice 5 with 90 probability Then the traveler must arrive at the bottleneck by the end of slice 3 Now suppose that a Type 3 traveler desires to depart at the end of slice 5 and that its earliness cost is 5 hour its queueing cost is 10 hour and its lateness cost is 20 hour and that the road has no toll Then the optimal arrival time is found by comparing alternatives Arrival Earliness Lateness Travel Total Time cost cost cost cost 5 min 18 60 5 0 2 60 UD 1 83 10 min 12 60 5 0 3 60 10 1 50 15 min 5 60 5 0 5 60 10 1 25 20 min 0 3 60 2 8 60 10 2 33 25 min 0 10 60 20 10 60 10 5 00 For the example an arrival time of 15 min the end of slice 3 is preferred BTS assumes that only a portion of the travelers would own the reguired eguipment to access the automated route SO each of the three classes is divided into two categories designated automation eguipped AE and not eguipped NE The latter group chooses the optimal route among those that are accessible B 2 Route Assignment rn O illustrate how vehicles are assigned to routes suppose that 50 of vehicles fall in the class AE and that 20 of vehicles are allowed to
16. fects route choices for Type 3 travelers Toll can also be specific to traveler type but the current version of BTS does not use tolls in selecting routes for Type 1 and Type 2 travelers B System Breakdown O Duration of breakdown l l Range 1 00 12 00 Dimension scalar Unit time slice Two parameters provide lower and upper bounds of a uniform probability distribution used to simulate incident duration DO Capacity loss distribution Range 0 00 1 00 Dimension scalar Unit none The level of the capacity loss is expressed as a proportion in 10 increments The sum of the proportion must equal one C Driver Choice Behavior U Arrival multiplier per iteration Range 0 00 4 00 Dimension scalar Unit none The multiplier by which traffic is incremented after each iteration The value is applied identically to each traveler type and each route at the end of each iteration after accounting for all shifts in arrival time and route For example suppose that 400 Type 2 travelers are assigned to time slice 5 and Route 2 and that the multiplier is set to 1 1 Then the multiplier will increase the number of travelers to 440 at the start of the next iteration Q Type 3 traveler costs Range 0 00 99999 00 Dimension scalar Unit dollar s hour Three parameters account for personal cost with respect to earliness lateness and queueing time per unit time for Type 3 travelers cost minimize
17. ime data These adjustments can include a demand increment to account for a general increase in traffic from year to year as well as changes in route and arrival time at the bottleneck BTS can run up to 10 iterations representing 10 years Travel time performance is automatically stored after each iteration for later analysis The following two sections explain the key phases of BTS the performance evaluation step and the demand assignment step A Performance Evaluation Each iteration of BTS simulates traffic conditions over a period of up to 200 days Each day will be called a run See Table 1 1 For each day the time duration and magnitude of freeway Time Slice A time increment of five minutes during which arrival rates and capacities are assumed to stay constant Run A single day comprising up to 45 consecutive time slices 3 3 4 hours Iteration A set of runs during which arrival time and route assignments as well as traffic volumes do not change An iteration might represent one year Table 1 1 Definitions incidents are randomly simulated according to user specified probability distributions The purpose of the simulation is to represent actual bottleneck capacity within each time slice ona given day taking into account the incidents that occur on that day A user specified probability determines the likelihood that an incident occurs in any time slice If an incident occurs
18. leneck operates below capacity On any day of the simulation the total time spent waiting in queue measured in hours is found from a summation of the queue sizes N Total Time in Queue W o O 1 n i l The average system delay is found by averaging W over all of the days simulated Because the simulation includes incidents the average system delay is the sum of the average recurrent delay and the average non recurrent delay To calculate average recurrent delay BTS simulates one additional day in which no incidents occur The output from BTS provides both this value and the average system delay on a per vehicle basis The difference between the values is the non recurrent delay In addition BTS provides the average free flow travel time which can vary when travelers shift between routes having different travel times BTS has the capacity to analyze up to three parallel routes Two of these routes are accessible to all travelers while the third is designated as automated The automated route is only accessible to a user specified portion of travelers that owns necessary equipment BTS provides system delay recurrent delay and free flow travel time for each of the three routes If desired the third route might alternatively represent an HOV lane accessible to a user specified portion of travelers As input to the traffic assignment phase sy
19. lso used to format output O Proportion of vehicles equipped with automation Range 0 00 1 00 Dimension scalar Unit none The proportion of the vehicles initially assigned to Route 1 and 2 that own the reguired eguipment to access Route 3 the automated route 100 of the vehicles initially assigned to Route 3 are automated D Traffic Volume Distribution O Number of travelers in slice i Range 0 00 99999 00 Dimension 1 45 Unit vehicles Initial assignment of travelers to time slices The volume is 14 the sum among all vehicle types and routes and does not distinguish between a desired arrival time and a desired departure time Let Aj volume in time slice i Then if Ag is set to 500 a total of 500 travelers desire to depart or arrive from the bottleneck at time slice 6 Each time slice represents a five minute interval E Simulation Control Data The following data can only be changed after selecting Run Simulation Program Number of runs per iteration Range 1 200 Dimension scalar Unit none Specifies the number of runs in each iteration Each run is gt equivalent to a single day of up to 45 5 minute time slices or a maximum period of 3 hours and 45 minutes O Number of iterations Range 1 10 Dimension scalar Unit none Specifies the number of iterations per simulation Each iteration represents a block of time e g a year over which rout
20. rs 13 QO Initial routing Range 0 00 1 00 Dimensions 1 3 1 3 Unit none Nine parameters determine the initial proportional assignment of travelers to route and type Let Ri j Proportion of travelers that are type i and choose route j Ri j is multiplied by the total traffic volume specified earlier to obtain the travelers by type and route As an example suppose that 400 travelers are assigned to time slice 6 with R 0 2 R 0 2 and R23 0 6 160 are Type 1 travelers and 240 are Type 2 travelers All Type 2 travelers initially choose Route 3 whereas half of the Type 1 travelers choose Route 1 and the other half choose Route 2 The sum of the proportions must equal one As the simulation progresses travelers are allowed to change route However traveler type stays the same in all iterations UO Proportion of the vehicles that would change route Range 0 00 1 00 Dimension 1 3 Unit none For each route a parameter specifies the proportion of travelers that will shift to an alternative route after an iteration if a lower cost route is available The same proportion is used for all traveler types O Percentile queue for Type 2 travelers Range 0 00 1 00 Dimension scalar Unit none The percentile of the waiting time distribution used by Type 2 travelers in selecting arrival time and route The probability that Type 2 travelers depart on time The parameter is a
21. s on the screen or print them out by selecting Display Output File in the main menu 17 The output file Report prn contains all of the input data as well as the summary statistics shown previously A more detailed output is provided in the file Lotus prn It contains the recurrent delay nonrecurrent delay and a user specified percentile delay by time slice and by route a cumulative traffic volume by time slice and by route an upstream delay by time slice and a total arrival volume by time slice This file can also be examined in Lotus 1 2 3 as discussed in the following chapter 18 CHAPTER 5 I EVALUATION OF BOTTLENECK PERFORMANCE WITH LOTUS 1 2 3 After the simulation is complete system performance can be evaluated in greater detail using Lotus 1 2 3 This can be done in the following steps 1 Return to DOS and enter your LOTUS 1 2 3 program 2 Select the File Retrieve option and select Lotshell wk1 from the drive that contains your BTS program 3 Move the cursor to the cell that states Import Lotus prn here 4 Select the File Import Numbers option and select Lotus prn from the drive that contains your BTS program You will now have access to graphs and data that provide arrivals and queueing time by route time slice and iteration Each column is formatted as follo
22. stem travel time per vehicle is also calculated by time slice as a function of arrival time at the bottleneck System travel time per vehicle is found by dividing the queue size by the actual capacity for the given day and adding the free flow time BTS averages these waiting times among the days simulated to determine an average system waiting time by time slice BTS also determines waiting time for a user specified percentile of the waiting time distribution A final feature is that BTS allows the user to specify a roadway capacity upstream from the bottleneck For BTS to operate properly this capacity must be larger than the nominal bottleneck capacity In the output BTS shows the recurrent delay upstream This represents the minimum possible delay on the highway should the bottleneck be removed without making ancillary improvements upstream B Traffic Assignment At the end of each iteration of BTS waiting time data is used to revise the assignment of vehicles to arrival times and to routes BTS allows for three types of traveler behavior QO Fixed arrival time These travelers always arrive at the same time independent of queueing However they will shift between routes to select the fastest available counting both queue time and free flow time O Fixed departure time These travelers would like to depart from the bottleneck at a time that insures they will arrive at their des
23. tination at a fixed time with high probability Travelers choose the route which allows them to leave home as late as possible and still arrive at their destination on time with high probability QO cost minimizers These travelers have a desired time to depart from the bottleneck but are willing to depart either earlier or later if their personal cost is reduced Travelers choose the route that offers the minimum average cost which is the sum of an early cost a late cost a travel time cost anda roadway toll if applicable Based on the above criteria as well as waiting time and free flow travel time data BTs determines the optimal route and arrival time for each type of traveler Travelers are automatically reassigned to their optimal arrival time However only a user specified proportion of travelers is allowed to shift between routes at any iteration Reassignment only takes place at the end of an iteration B l Arrival Time Assignment As an illustration of the arrival time assignment suppose that the following travel time data is available for a route The travel time will represent the sum of the queue time and free flow travel time Time Slice Elapsed Time Mean Travel Time 90 Travel Time End of Slice End of Slice End of Slice 1 5 min 2 min 3 min 2 10 min 3 min 5 min 3 15 min 5 min 10 min 4 20 min 8 min 17 min 5 25 min 10 min 21 min Because Type 1 travel
24. ws Column A Time Slice Column B Cumulative Arrivals Route 1 Column C Percentile Wait per Vehicle Column D Average Wait per Vehicle Column E Recurrent Wait per Vehicle Column F Cumulative Arrivals Route 2 Column G Percentile Wait per Vehicle Column H Average Wait per Vehicle Column I Recurrent Wait per Vehicle Column J Cumulative Arrivals Route 3 Column K Percentile Wait per Vehicle Column L Average Wait per Vehicle Column M Recurrent Wait per Vehicle Column N Recurrent Wait per Vehicle Upstream Column 0 Cumulative Arrivals Sum of Rte 1 2 3 Column P Percentile Wait per Vehicle Column Q Average Wait per Vehicle Column R Recurrent Wait per Vehicle The data can be analyzed through any of the regular Lotus 1 2 3 features or through the following predefined graphs To display a graph enter the commands Graph Name Use followed by the graph name To print a graph first save it by entering the commands Graph Save followed by a graph name The graph can later be printed from the Printgraph option of Lotus 1 2 3 19 The following are the preset graphs contained in the Lotshell wkl file Name Function CUMRT1 Cumulative arrival curve Route 1 iterations 2 4 6 8 CUMRT2 Cumulative arrival curve Route 2 iterations 2 4 6 8 2 4 6 8 2 4 6 CUMRT3 Cumulative arrival curve Route 3 iterations CUMTOT Cumulative arrival curve total all routes iter
Download Pdf Manuals
Related Search