Home

Novel Order Book Rules to Promote Trading System

Contents

1. Creates a new Java SE application in a standard IDE project You can also generate a main class in the project Standard projects use am IDE generated Ant build script to build run and debug your project 2 In the next window give the project a name OrderBookSimulation which by default will also create a sub folder OrderBookSimulation in your D dissertation folder Select the option Set as Main Project while not selecting the option Use Dedicated Folder for Storing Libraries Click Finish Steps 1 Choose Project OrderBookSimulation 2 Name and Location EI XR Eee D Missertetion 46 Lingcen Huang fh Novel Order Book Rules to Promote Trading System Price Stability 3 In the main project window right click on project OrderBookSimulation and select Properties In the next window choose the category Sources and add the folder D dissertation mason under Source Package Folders while not adding any folders under Test Package Folders Then in the same window choose the category Libraries and under the tab Compile click on the option Add JAR Folder and select all JAR files available in folder D Missertationwnason Click OK B Run 9 Application O9 Web Start Formatting F 11 i 9 Compiling O Packaging 9 Documenting o Run B o Application 079 Heb Start i Q Formatting le D dissertation mason itext 1 2 jar a D dissertation mason j3
2. 1 Legend i E Antialias C ENENS 1 MM Ovi 50 Lingcen Huang ih Novel Order Book Rules to Promote Trading System Price Stability 8 Click the Console tap to change simulation speed and re set random number Delay SeciStep Q 00 Thread Priority 7 Steps per Step Button Q 1 Automatically Stop at Step Automatically Stop after Time Automaticalty Pause at Step Automatically Pause After Time Random Number Seed 1315343065270 Increment Seed on Stop Repeat Play on Stop 9 Close all pop up windows to finish simulation Test 1 Go to NetBeans main menu and select Run button Single click second line Test Project OrderBookSimulation T java uthUI java Order Interface java yer java rer java Stop wn java 51 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability 2 Wait for 8 seconds and test result will display in bottom window The test result is displayed in below inib Deleling D disseclalion OrderBookSigulalion build buill jac proper Lies deps jar Updaling properly File D disserlaLion OcdecBookSimulatlion build buill jarc proper lies inil deps rlsan Updaling properly Tile D XdisserlaLionNOrderBopkSiaulalLionhbuildhbuill clean properlies Deleling diceclory D disserlaLion OrdecBookSigulalion build clean Crealed dic D disserlaLion OcdecBookSiaulalion build
3. For scenario 1 go to ModelSetting class to modify code in Properties properties new Properties try properties load new FileInputStream setups CBook properties For scenario 2 go to ModelSetting class to modify code in Properties properties new Properties try properties load new FileInputStream setups DA Book properties For each time the parameters setting are different We use parameter D to control maximum and minimum threshold for trading The maximum volatility is up to 0 3 and minimum volatility cannot be less than 0 003 The details of each parameter setting for each time and each scenario will be shown in table 8 and table 9 Scenario 1 Cont Book Model Run Times Parameter D Other Parameters Ist 0 3 maxT 10000 a 0 675 6 0 29 o 1 u 0 525 2nd 0 03 initial price 12 0 ContPlayer 1500 3rd 0 003 Table 8 Scenario 1 Parameters Setting Scenario 2 Double Auction Book Model Run Times Parameter D Other Parameters Ist 0 3 maxT 10000 a 0 675 6 0 29 o 1 u 0 525 2nd 0 03 initial price 12 0 PatientPlyer 900 MixPlayer 3rd 0 003 800 Table 9 Scenario 2 Parameters Setting 34 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability 5 2 Scenario 1 Cont Book Model This model is simple one without any order book rules All traders follow the same trading strategy T
4. The project needs agents having a spatial component to their behaviors and interactions The past of the market is no predictor of the future Scaling up to arbitrary levels is important Process structural change needs to be a result of the model rather than a model input 1 2 Aim and Objectives The main aim of the project is to find out the novel limit order book rules to improve the trading price stability The objectives in order to achieve this aim are To do overview of the existing limit order book models And to analyze the issues related to building an effective limit order book To build a simple order book model without any trading rules To analyze the price movements for the simple order book model To develop new algorithms to limit the behaviors of the order book and to introduce them into the simple model to extend model capabilities To analyze the price movements having applied the order book algorithms To compare the results output from the simple model and the extended model and discuss the improvements in price stability 5 Lingcen Huang Ier Novel Order Book Rules to Promote Trading System Price Stability The objectives that must be achieved in the design order book model are Three agents must be built to present the order book a buy trader and a sell trader Order book and traders must communicate with each other The order book model must apply effective algorithms The s
5. et al Preis 2006 as a multi agent based order book model of financial markets They introduce a simple model for simulating financial markets based on an order book in which several agents trade a kind of stock at a stock exchange continuously For a fixed market structure of model the used price time priority matching algorithm produces a diffusive behavior of price and order flow rates of the different kind of order types They also show a market trend for example an asymmetric order flow of any type can result in a non trivial Hurst exponent for the price development but not fat tailed return distributions When an extra couples the order entry depth to the prevailing trend our Order Book Model can also reproduce the stylized empirical fact of fat tails Another worthy research is carried out by Biails et al Biails et al 1995 who do an empirical analysis of the limit order book and the order flow in the Paris Bourse They analyze the demand and supply of liquidity For instance thick books result in trades while thin books elicit orders In order to get time and price priority investors quickly place orders within the quotes when the spread is large or depth at the quotes 16 Lingcen Huang rea Novel Order Book Rules to Promote Trading System Price Stability 3 2 Use Cases and User Requirements Analysis 3 2 1 Use Cases Analysis Before analyzing the user requirements and user cases who will use this sy
6. London 23 Preis T 2006 Multi agent based Order Book Model of financial markets PhD Thesis Germany Johannes Gutenberg University of Mainz Staudinger Weg 24 Rama Cont 2006 Long Memory in Economics chapter Volatility Clustering in Financial Markets Empirical Facts and Agent Based Models Springer Berlin Heidelberg 25 Ranaldo A 2004 Order aggressiveness in limit order book markets Journal of Financial Markets 7 53 74 26 Rosu I 2009 A Dynamic Model of the Limit Order Book Review of Financial Studies 22 4601 4641 27 Sanmay D 2008 The Effects of Market Making on Price Dynamics Proceeding AAMAS 08 Proceedings of the 7th international joint conference on Autonomous agents and multiagent systems 2 24 32 28 Whigham P A 2006 Evolving trading strategies for a limit order book generator In Proceedings of the Congress on Evolutionary Computation CEC Barcelona 18 23 July New Zealand IEEE Press 1 8 cannot be both in Barcelona and New Zealand select the correct one 29 Withanawasam R Whigham P Crack T and Premachandra I 2010 An empiricial investigation of the Maslov limit order market model Information Science Dept Univ of Otago Tech Rep 2010 4 58 Lingcen Huang
7. Save Model The developer saves the existing model with a special name into hardware for users loading to use The user selects the model or type the model name to accept 3 2 2 User Requirements Analysis Requirements gathering and analysis as the important step in the project are significant tasks for further understanding of the system s behavior and the project scope It is critical to the success of the development process of the simulation model They are high level functions that user wish the system can provide They should not include the technical details but they should provide enough information to make developer decide whether to apply those functions into the model UR 1 The system must provide model creation utilities UR 2 The system must provide simple user interface UR 3 The system must allow user change the parameter setting before the simulation run or during the simulation run period UR 4 The system must provide a visualization window for the current state of the simulation and animate this during simulation runs UR 5 The system must allow user to review all ask price and bid price together or separately UR 6 The system must allow user to review all ask depth and bid depth together or separately UR 7 The system must allow user to export and save the statistical results into text files 19 Lingcen Huang wet Novel Order Book Rules to Promote Trading System Price Stability 3 3 Sys
8. uoe ste hiu uonmusura duy Ay pue aduanbas wesbeig ssel weage u6isag wass 21589 ufisag pawalig 139 90 SISABUY wawasinbay waiss sis euy 585 Ja5f sis jeuy payaug pafqo 85 a5f fuuaywen syuaueJimbany sase asf pue uawearinbay Malady amena pniS aJe 10S VOLH 3014 12 014 3J0UJA OWEN a 12 02 6T 8T il oT ST PI a a ni o m OO Lingcen Huang 53 ET Novel Order Book Rules to Promote Trading System Price Stability Project Schedule Week Description Finished or Not 1 II Project Group Meeting Finished Introduction to Mason Think about project topic 2 Review Mason tutorial Finished Choose project topic Write objectives 3 Install the relevant software Finished Set up 4 Run tutorial program under instruction Finished 5 Literature review Finished Trading strategy Multi agent trading system model 6 Project plan and Gantt chart Finished 2 project group meeting 1 individual meeting 7 Object oriented analysis Finished 2 individual meeting 8 Object oriented design Finished 3 individual meeting 9 Implementation Finished 4 individual meeting 10 Implementation Finished Test 5 individual meeting 11 Implementation Finished Results analysis Write dissertation 12 Write dissertation Finished 13 Write dissertation Finished 14 Write dissertation Finished 54 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability Appendix D Sourc
9. 04 html 13 Jiang J 2009 Revealing Intraday Market Efficiency Estimating Diurnal Price Densities in Limit Order Books In ICIFE 09 Proceedings of the 2009 International Conference on Information and Financial Engineering Washington DC USA IEEE Computer Society 8 12 14 Kissell R Malanut R 2006 Algorithmic decision making framework Journal of Trading 1 1 1221 15 Lamba H 2008 Rational expectations psychology and inductive learning via moving thresholds Physica A 387 3904 3909 16 London Stock Exchange LSE 2011 Rules of the London Stock Exchange Effectivel 35 43 17 Luke S 2011 Multi agent Simulation and the MASON Library PhD Thesis USA George Mason University 18 Macal Charles M 2006 Tutorial On Agent based Modeling and Simulation Part 2 How to Model with Agents In Proceedings of the 2006 Winter Simulation Conference U S A IEEE Press 73 91 19 Marco A 2007 High frequency trading in a limit order book Quantitative Finance Vol 8 217 224 20 Market model website http svn2 assembla com svn MarketModel 57 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability 21 Maslov S 2000 Simple model of a limit order driven market Physica A vol 278 pp 571 578 22 Manyumwa E 2010 A Multi Agent Simulation of an Intraday Liquidity Market With the Use of Prediction Markets M Sc London University College
10. 6 Chapter 2 Background and Research Information eee 7 Bel Order Bokss aeiiao 7 2 1 1 Order Book Definition a sce cn eds omoia bin tur pcs Sae RR EN E Ult PER CO RUE 7 2 1 2 Limit Order Book and Model cisecie te ta Babe to Blot Serv tuus preuve pi Des 8 2 2 Order Book Trading Rules and Strategies eee 10 2 2 1 Order Book Trading RUNES 22st na ER Ur i tor RH nee 10 2 2 2 Limit Order Book Strategies ius rit teen rb sa da IR Hd S ERI eee el Koen repas 12 2 3 Agent based Modeling eene ceti cute kno kho nein eb H VE PVP FEY V TOP SEVERE EUR tke EY CRUCE Td 14 Chapter 3 Order Book System DOeSIgu uoisieesvs se eibi suit pen ta qune RE EkEP EE Su IP eS RB pPN n Yep 16 3 1 Guidance from Related Research eee eee eee eee enne enn 16 3 2 Use Cases and User Requirements Analysis eee e eres eere eee nne 17 925 Use Cases A MALY GIB NITET PO 17 3 2 2 User R quirements ATI VAIS saccacascusheavidssssaiunennatsadeuvscnadsabeisenvertucenseciaeess 19 3 3 System Requirements me eT 20 3 4 The Order Book Model Desigitscccsicsscsscsnssccosssssenvssnsecovsnsssdenseioancocnaccecunsiecnanse 21 Chapter 4 Implementation and Testing e eeeeee eee eee e entente nete nete natn annu 26 4 1 Software Application Used ue neci nox ob t En Pe Ok vVAF CI URP PE PALO E A GS 26 4 2 System Implementation eas cecei coner ido teer bd ote p EPA po Ua E
11. classes Crealed dic D disserlalion OcderBookSigulaLlion build eaply Crealed dic D disserlalion OrdecBookSigulalion build generaled sources ap source puLpul Compiling 391 source Files Lo D disserlaLion OcderBookSiaulalion build classes HR AES XE Ema RT Bits APT TE BTR WKAR XlinL deprecalion AAW HR ACRES X EAE T ALERT SE TEE STAGE WKAR XlinL unchecked S rs Copying 508 files Lp D disserlaLion OrderBookSigulaLion build classes compile compile Lest Lesl reporl lest BUILD SUCCESSFUL Lolal Lime 8 seconds 52 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability t Plan Projec Appendix C Gantt Chart 1102 60 61 T102 60 T T102 80 St T10z 80 TT 1102 80 60 1102 80 50 1102 80 60 1102 80 0 1102 20 62 Todi olet 1102 80 10 T102 z0 ST 1102 20 ET 1102 20 81 1102 20 80 1102 20 50 1107 20 80 1102 90 06 1102 90 91 1102 60 1 on om jwn wr on ooze urit owe a aw jag js ws ort foso ug TI quads 102 asnbing Toe Anf 1102 60 1 1102 80 51 TTOZ B0 TT 1102 80 60 1102 80 50 1102 80 0 1102 80 0 1102 80 10 1102 20 52 1102 20 81 1102 20 81 Troz z0 T TTOZIZO TT 1102 20 11 1102 20 50 1102 20 10 1102 20 10 1102 90 21 T102 90 1 1107 90 1 Wei POT pod pot pot pot pot pot pov pos pos pot pot pot pos pov pot po p oor pov p 0 89 uongung UOISSNUQnc UONELASSIC IIUM SISARUY 1jn58y 158 ndino pue indu
12. for liquidity provision So it is necessary to understand the limit placement the contribution to liquidity and price formation Jiang 2009 We also are very interested in the dynamics of the order book and order flow the order flow not only reacts to the information from the market and the order book state but also influences the trading activity in the marketplace affects the evolution of the order book afterwards and updates the book state As known an order book has extremely rich data and complex behavior Therefore it is important to model and study its behaviors and develop algorithms to limit its behaviors It will be helpful in improving price stability and reducing trading costs 1 1 2 Agent based Modeling Motivation Agent based modeling is a novel way to build a system model that is comprised of interacting autonomous agents Macal 2006 ABM could have a great positive effect on the way that scientists do research with electronic labs or businessmen use computers to make decisions We can take all types of independent components as agents The agent behaviors in this project can range from simple ordering algorithm to complex limit order book matching rules From this project model standpoint the following agent characteristics will be relevant Macal 2006 3 Lingcen Huang ere Novel Order Book Rules to Promote Trading System Price Stability 1 Identifiable the rules can be set to limit its behavi
13. is for cont book model The number of trader can be changed in txt files All parameters influenced simulation results in set up files are described in table 7 Parameter Symbol Explanation maxT Max number of ticks D Standard deviation of the noise numAssets Number of assets present on market initialPrice Initial price Cont_lambda Cont market depth Cont_s Cont ratio of traders updating threshold Farmer_alpha Patient trader orders placed per step Farmer_delta Patient trader orders canceled per step Farmer_mu Impatient trader orders placed per step Farmer_sigma Shares per trade order size Table 7 Explanations of Parameters 31 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability 4 3 Testing 4 3 1 Testing Process After finish implementation system the next step must be done is testing The proposal of testing is to make sure system operate normally work as expected and meet all functional non functional requirements that guide it design In this project testing is done exactly in the process that shown in Fig 14 Make Testing Plan Build Test Enviroment Design Build Test Procedures Execute System Test Execute Acceptence Test Review Anaylsis Figure 14 Testing Process Diagram Design System Test Making Testing Plan Creates test schedule plan and test approaches Design System Test Identifies test c
14. more complex strategy the limit price can be set according to market condition dynamically Daniel 2006 For instance if a trader wants to make profit and minimize the cost it will be a good choice to place passive orders when the price is decreasing and to submit aggressive orders when the price is rising up In Peter A Whigham s paper a limit order book generator model is developed and it is tested whether a grammar based evolution GE can evolve a strategy to make profit or not Whigham 2010 Grammatical evolution is an evolutionary automatic programming technique to help seed the evolutionary process with the trading strategy The model used to generate a limit order book is called the Maslov model Maslov 2000 For this model the trader is chosen randomly at each time step and he may do a limit order or a market order The number of stock sold or bought is not limited Fig 3 shows the algorithm for Maslov generator to generate a limit order book L in a single time step Maslov 2000 L Price is last trade price L Se and L Buy are limit 13 Lingcen Huang ih Novel Order Book Rules to Promote Trading System Price Stability order sell buy books Z Sell Price and L Buy Price are current sell buy top entries in books Input Limit Order Book L Tick Size t LimitOrderProbability plo Output Updated Limit Order Book L buy or sell if Rnd 0 1 2 0 5 Buy behaviour if Empty L Sell orRnd 0 1 splo in
15. thresholds in light of current volatility by setting it to the trailing period s return rate A Better User Interface The user interface is still needed to do further development When run the simulation model an interface window for setting parameters should be displayed before simulation charts output It will be much more convenient for changing set up files every times before run An interface for select model also should be done instead of modifying codes every times Data Output and Storage In this project the model only can output results in charts style The further work is to develop function that all simulation data can be store in a file as another results output The data could be saved in txt file But because of the huge mount of data a more efficient approach for storing data is necessary This approach should use less memory and is easy to manage data for users 44 Lingcen Huang era Novel Order Book Rules to Promote Trading System Price Stability Appendix A System Manual Computer System Requirements Windows XP Vista 7 512 MB RAM Minimum 1GB RAM recommended Hard Disk 200 MB for NetBeans 7 0 1 DVD Rom Drive Software Requirements JDK 6 Update 27 with NetBeans 7 0 1 Mason Java 3D 1 5 1 System Set Up Steps 1 Download the file mason zip from http cs gmu edu eclab projects mason 2 Create a folder dissertation in your D drive 3 Extract mason zip directly into disser
16. ASON Framework Ph D Center for Social Complexity George Mason University USA 7 Cui W 2010 Evolving Efficient Limit Order Strategy using Grammatical Evolution In 2010 IEEE World Congress on Computational Intelligence Barcelona U S A IEEE Press 2408 2413 8 Cui W Brabazon A O Neill M 2010 Evolving dynamic trade execution strategies using grammatical evolution in Proceedings of EvoFin 2010 Applications of Evolutionary Computation Lecture Notes in Computer Science 6025 Springer Verlag Berlin 191 200 9 Daniel G 2006 Asynchronous simulations of a limit order book Ph D Thesis University of Manchester U K 10 Ghoulmi F 2007 Effects of diversification among assets in an agent based market model In Proceedings of the Complex Systems II Conference Australian National University 4 7 December ACT Australia Canberra 12 56 Lingcen Huang eru Novel Order Book Rules to Promote Trading System Price Stability 11 Farmer J Patelli P and I Zovko I 2005 The Predictive Power of Zero Intelligence in Financial Markets ArXiv Condensed Matter September 2003 Sean Luke Claudio Cioffi Revilla Liviu Panait Keith Sullivan and Gabriel Balan MASON A Multiagent Simulation Environment Simulation 81 517 525 12 Frank H Westerhoff Multiasset Market Dynamics Macroeconomic Dynamics 8 05 596 616 November 2004 URL http ideas repec org a cup macdyn v8y2004105p596 616
17. University College London MSc Financial Computing Project 2010 2011 Supervisors Antoaneta Serguieva Chris Clack Submission date 14 September 2011 This report is submitted as part requirement for the MSc Degree in Financial Computing at University College London It is the result of my own work except where explicitly indicated in the text The report may be freely copied and distributed provided the source is explicitly acknowledged but the real data used is confidential THE DATA MAY NOT BE COPIED OR DISTRIBUTED EXCEPT WITH PERMISSION FROM THE PROJECT HOST ORGANISATION wes Novel Order Book Rules to Promote Trading System Price Stability Abstract This project focuses on designing an agent based order book simulation model that applies novel rules to improve market price stability The model is developed by Java language To implement this model Mason software and some java libraries are also applied In order to compare the difference of price fluctuation when apply and not apply those order book rules we develop a simplest model firstly This model has an order book without any rules and all traders have same behaviors And then we improve this model by change simplest order book to double auction order book The traders can be patient players or impatient players The patient players place limit orders which include the price and quantity they want to sell or buy The impatient players place market orders which execute immediate
18. ackage has two sub packages agents and market It use GUlIdisplay package to generate graphics The description of each package and package relationship diagram are show in following Model Agents Test uses a j Model Market i I i LI uses lt lt uses gt gt GUldisplay I I Figure 12 System Package Diagram D Lingcen Huang Ire Novel Order Book Rules to Promote Trading System Price Stability GUldisplay It includes a SimulationModelWithUI to plan chart style and generate the visual outputs DisplaySupporter applied SimState simulation engine to load simulation scenarios and build GUI An html file gives a description of simulation model for users Model It includes two classes that builds a simulation model structure and controls setting Two sub systems Agents and Market are belonging to this package Model Agents It includes three different types of trades ContPlayer MixPlayer and FarmerPatientPlayer A class defined as interface to connect with simulation model Model Market It includes a class that is used to simulate the trading environment It can generate price for market It also includes two order books One is simplest order book another is limit order book with some rules An interface class is created in this package to do the communication between order book and market Test It includes a test class to test the system whether can work with n
19. ade components aimed to improve price stability and reduce transaction cost The components include how many orders should be submitted size for each order type of each order and what time should be submitted to market The trading volume of an order could be presented by a percentage of the average daily volume of the stock If the volume of an order is less than 5 of average daily volume it can be traded without any complex strategies Kissell 2006 If the volume is more than 159 it will lead to a high market impact In order to reduce the impact the order should be executed in several trading days The normal size of an order 1s 596 to 15 of average daily volume An appropriate trading strategy should be applied to execute these normal size orders The reason for applying a trading strategy when an order volume is more than 5 of average daily volume is that it has the potential to change the price of that asset When price changes there will be a huge market impact cost If a large order could be divided the cost will be reduced On the other side suffering opportunity cost could be increased when large orders are divided Therefore a good trading strategy should try to balance these costs and reduce the total cost as much as possible In the next paragraphs several strategies will be discussed that have been outlined in published research papers The simplest one is a pure market strategy Cui 2010 An order that i
20. and the bid price in M green line 7 The model shall output results of the limit order price in graphs 8 The model shall provide an interface to modify the parameters such as time scale etc 9 The model shall implement the trading strategies for trader C 10 The model shall allow simulation results history saved automatically C Table 5 Functional System Requirements 20 Lingcen Huang rad Novel Order Book Rules to Promote Trading System Price Stability ID Non Functional Requirement Priority 1 The model shall allow users to import native java libraries or external S libraries to extend its functionality 2 The model shall be extensible allowing users to change functionality C without changing many codes Table 6 Non Functional System Requirements 3 4 The Order Book Model Design In modern electronic exchanges the order book saves demands and offers of different traders and supports a continuous trading The minimum price change is called tick size The price can be calculated by multiple of the tick size The lowest offer price is called best ask while the highest demand is best bid The spread is non zero gap between best ask and best bid as shown in Fig 6 Limit orders at any price level will be added chronologically in the queue of that price tick according to the price of the limit orders Then a price time priority matching algorithm could be ach
21. ases test cycles exit amp entry criteria and expected results Design Build System Test Procedures Sets up procedures like status reporting and error management Build Test Environment Request and builds data set up software and hardware Execute System Test Tests system performance and reports errors Execute Acceptance Test Tests users acceptance Review Analysis of Results Analyzes and summarizes tests results in report documents S Lingcen Huang ra Novel Order Book Rules to Promote Trading System Price Stability 4 3 2 Scope Approach and Results Scope 1 Functional Validation Testing Objective Make sure system meet functional requirements from design specification Approach Run the system for 10 times with different settings and approaches Compare the functions achieved actually by system with the requirements in design specifications Do intensive testing of the new front end fields and screens windows GUI standards screen amp field look and appearance and overall consistency with the rest of the application Results The system can build order book and match orders correctly The parameters can be set in set up files and be read in correctly Simulation graphics is displayed as expected Scope 2 Technical Testing Objectives Make sure all unit java class run successfully and links between classes are correct Approach A test package is developed in system In the package a specific class calle
22. cs in the interface windows View Graphics The user chooses to see the output in a graphic way The model should show the result visually by graphics Export Results The user exports and saves the statistical results in a text file The user exports and saves the graphic results in a jpg or gif file Run Simulation The user changes the simulation parameters optionally and clicks the start button to run the simulation The system begins the simulation and provides visual feedback while it is running The user chooses to pause or stop the simulation at any time The system stops running the simulation and saving the model s current state allowing the user to run simulation continuously from where it was stopped Set Parameters The developer sets the initialization value of parameters in the system The users choose to change the value of parameters in the text file The system will save new values of changed parameters Apply Algorithms The developer applies the matching algorithm for the order book and trading algorithms for traders in developed simulation model Test Model The developer calibrates validates the simulation model in the computer system Edit Model The developer edits and modifies the existing simulation model 18 Lingcen Huang Bee Novel Order Book Rules to Promote Trading System Price Stability Create Model The developer designs the simulation model and implements it into the computer system
23. d TestValidation will test unit and links of classes Click Run button and select test project tap to run test code The code in details will be list in appendix Results After run test class netbeans reports 0 error found Scope 3 Performance Testing Objective Make sure the system provide acceptable response times Approach Run the system without GUI for 10 times and run the system with GUI for 10 times Results Without GUI the system response time is 3 seconds With GUI to view all charts 20k steps need to run The response time is up to 3000 seconds Scope 4 User Acceptance Testing Objective Make sure the system operates in the manner excepted Approach 25 test cases were performed for all supporting material like form and procedures are suitable and accurate for the purpose intended Results There is no any error found during acceptance testing 20 of 25 cases have been completed successfully 33 Lingcen Huang lores Novel Order Book Rules to Promote Trading System Price Stability Chapter 5 Results and Analysis This chapter analyzes the experimental results This chapter displays all resulting output from the system These results are compared and analyzed The findings are also discussed in this chapter 5 1 Simulation Plan The simulation will run with 3 times for two scenarios cont book model and double auction book model To alert scenarios the code need be modified to import set up files
24. d close the file 7 Run simulations After run this model and collect the results we change model setting 8 Go to ModelSetting java class and active the command Properties properties new Properties try properties load new FileInputStream setups DA Book properties PA 9 Open DABook file in OrderBookSimulation folder and set the value of a 6 o u according to table 9 10 Save and close DABook 11 Open a txt file called farmercont in same folder 12 Set MixPlayer and FarmerPatientPlyer number according to table 9 save and close the file 48 Lingcen Huang ih Novel Order Book Rules to Promote Trading System Price Stability Run System 1 Go to NetBeans main menu and select Run button Single click first line Run Project OrderBookSimulation F6 Cdefault d AlttFG Test Project OrderBookSimulation p Build Project OrderBookSimulation Fil T Clean and Build Project rderBookSimulation ShifttF11 Set Project Configuration b fae Set Main Project gt r a Generate Javadoc OrderBookSimulation Order Run File Shi ft F6 Interface Orde Test File Ctrl F6 Compile File F9 ava Sp Check File ALt FS SE Velidete File Mttshieeerg 75 according er java Repeat Build Run OrderBookSimulation run lel are LOG pr r java Stop Build Run Byer extends a Output OrderBookSimulation run 2 The console window of the simulat
25. deore jar a D dissertation mason j3dutils jar D dissertation mason jcommon 1 0 0 jar in D dissertation mason jfreechart 1 0 1 jar D dissertation mason jmf jar ls D dissertation mason vecmath jar 4 Extract the Code zip from the DVD to a destination folder 5 Copy all sub folders in Code folder and pasty them into OrderBookSimulation folder 47 Lingcen Huang Rea Novel Order Book Rules to Promote Trading System Price Stability Set Up Simulation Model 1 Go to ModelSetting java class and active the command Properties properties new Properties try properties load new FileInputStream setups CBook properties pP EjfedelSetting java x SimwlationModel jsva sj TestValidation java S TradingEnviroment java lt B S AtSeBlretziguloul aa 26 27 28 29 if target null 30 returnSim true 31 target new SimulationModel System currentTimeWiliist 32 else 33 this target target 34 35 36 37 38 Properties properties new Properties 38 try 40 4l 42 43 catch lOException amp 2 Find the setups folder in OrderBookSimulation folder Open it and find a prosperities file called CBook 3 Open CBook file and set the value of a 5 o u according to table 8 4 Save and close CBook 5 Open a txt file called cont in same folder 6 Set ContPlayer number according to table 8 save an
26. e Code List Because a lot of codes are developed in this project the report will be too long if all of them are listed in appendix The appendix D will only list all packages class and other files names The detail code of each class will be found in DVD enclosed Package GUldisplay model model agents model market test Java Class BooklInterface ContBook ContPlayer DisplaySupport DoubleAuctionOrderBook FarmerPatientPlayer LimitOrder MixPlayer ModelSetting PlayerInterface RandomNum SimulationModel SimulationModelWithUI TestValidation TradingEnviroment Other Files index html ucl logo jpg 55 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability Appendix E Bibliography 1 Avellaneda M 2008 High frequency trading in a limit order book Quantitative Finance 8 217 224 2 Bartolozzi M 2010 A multi agent model for the limit order book dynamics The European Physic Journal B 78 265 273 3 Biais B 1995 An Empirical Analysis of the Limit Order Book and the Order Flow in the Paris Bourse The Journal of Finance 5 1665 1779 4 Black C 2010 Challenges for Equity Front Office Architecture M Sc Dissertation London University College London 5 Chen S 2007 Computationally intelligent agents in economics and finance Information Sciences 177 5 1153 1168 6 Class Project Team CSS 739 2009 Simulating Financial Markets using M
27. e bE E Rea EVEN EE CP TIE EN cH KEYS 27 4 2 1 Description of Java C 38S sscasicssncsssssenccsagecedesesastabessarsadestansncisansantesoiiacsnienss 28 4 2 2 Description of Simulation Model sse 31 LN EI i T PE TE A ESO EE EA E ET 32 4 3 1 Testing PROCESS Lus qaiccu ia run pH vx Ead e oce i E E AREE EES 32 4 3 2 Scope Approach and Results 15 2 5 Iter ty 98 enscvedinvasisevedeaesdenssenslienstudiss 33 Chapter 5 Results and ANalySiS ssssssssssssssissssessvsssssssosossosososssscsissssssissvssosssssvsssssiisso s 34 5 1 Simulation rium 34 5 2 Scenario 1 Cont Book Model cccccscccocssscssssccossscesccsseccsecccseccescessonsseeces 35 5 3 Scenario 2 Double Auction Book Model erect 38 5 4 Analysis of m A 42 Chapter 6 Conclusions and Future Work ssssccssscsssssssscssesssssssssssssesseesseseees 43 6 1 Project SUDAN Y eter r Q 43 MAG 43 6 3 Parr Work P 44 Appendix A System Manudl aiiis exon eti taroen e epbee nk KEERKR IUE bU Hu S E V pEU INFE EFE isses 45 Appendix D User Manual ones es iih o F R AER ERO ERE pa SER SUR U ix PORE FUB SER EE DH UIN HER RN S EE EUR 46 Appendix C Project Plat ua enisi a nad E DRUH BE LPIE REN DID adenosine 53 Appendix D Source Code List scccisccsssesssssssessasscncseessisntecssoasssaveasncvessoassoenevesdoascoseseas 55 Appe
28. e class for all order books This class provides the functions like placing an order for immediate execution creating a limit order and cancelling a limit order ContBook It is the simplest book that all traders follow the same behavioral rules DoubleAuctionBook This class implements a complex order book The market moves when market orders fulfill limit orders The limit order don t ever execute when they are placed LimitOrder This class contains the rules to improve the market price stability test package TestValidation This class is developed to test system validation 29 Lingcen Huang fh Novel Order Book Rules to Promote Trading System Price Stability This figures shows the whole java class structure of system Figure 13 Java Class Structure 30 Lingcen Huang Pill Novel Order Book Rules to Promote Trading System Price Stability 4 2 2 Description of Simulation Model This model is designed by Java language to simulate an order book system This system has different kinds of order books and traders Each trader has his own trading strategy A trading environment is simulated to place and execute the orders The data of trading results like price is output by GUI The GUI is achieved under help of Mason and java libraries The whole order book simulation system contains two scenarios One applies simplest order book model without any rules anther applies double auction book with limit rules The simples
29. e lowest ask and highest bid The difference between them is called the spread They are the signal that orders need to be fulfilled in the market Book depth It means at particular time in the book how many price levels are available Some books could have as many levels as wished while other books have a fixed depth i e the orders over the depth number will be rejected or ignored 7 Lingcen Huang Ire Novel Order Book Rules to Promote Trading System Price Stability 2 1 2 Limit Order Book and Model An order book that contains limit orders 1s a limit order book Black 2010 The Fig 2 shows the concept of limit order book in details The limit order book lists a current set of sell and buy offers for a particular instrument according to price levels The 1 entry in the limit order SELL book is the lowest price to sell the instruments and the other orders are listed in increasing price in the SELL book On the other side the 1 entry in the limit order BUY book is the highest price to buy instruments with other orders decreasing Chen 2007 In a limit order book the concept of last trade price is applied If a market order could match the top of the limit order book a trade will be done We can see the book in Fig 2 Last Traded Price Sort Descreasing Sort Increasing Limit Order Table Limit Order Table SELL BUY Figure 2 A Simple Order Book Trader A can sell a share of stock at price 50 best buy
30. ews all trades undertaken under its G rules as firm However the Exchange may in exceptional circumstances undertake an Exchange enforced cancellation of an automated trade executed on the trading system either at the request of a member firm or of its own volition In considering a member firm s request for an Exchange enforced cancellation the Exchange will have regard to a number of factors that are set out in the guidance below and whether 2120 1 both parties to the trade s are unable to agree to use the contra facility 2120 2 the request for an Exchange enforced cancellation is submitted to the Market Supervision department within a time period specified by the Exchange in the guidance to this rule 2120 3 the member firm requesting the Exchange enforced cancellation provides appropriate information to the Market Supervision department as set out in the guidance below and 2120 4 A member firm has incurred an amount of loss through an automated trade conducted on the trading system as specified in the guidance to this rule Table 4 Order Book Trading Rules of London Stock Exchange source London Stock Exchange LSE 2011 Rules of the London Stock Exchange Effectivel 35 43 11 Lingcen Huang Ire Novel Order Book Rules to Promote Trading System Price Stability 2 2 2 Limit Order Book Strategies Limit order book strategies is a set of rules to determine an amount of tr
31. his model will be run in three times and the parameters are set as details in table 7 The output charts include prices plots and average volatility plot The simulation results are shown as following Figure 15 Price for Cont Book Model at D 0 3 Figure 16 Average Volatility for Cont Book Model at D 0 3 35 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability Figure 17 Price for Cont Book Model at D 0 03 Figure 18 Average Volatility for Cont Book Model at D 0 03 36 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability Figure 19 Price for Cont Book Model at D 0 003 Figure 20 Average Volatility for Cont Book Model at D 0 003 37 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability 5 3 Scenario 2 Double Auction Book Model This model has two kinds of traders which trade a certain asset by the approach of a continuous double auction The traders are patient and impatient This model will be run in three times and the parameters are set as details in table 8 The output charts include prices plots average volatility plot and order book graphic The simulation results are shown as following Figure 21 Price for Double Auction Book Model at D 0 3 Figure 22 Average Volatility for Double Auction Book Model at D 0 3 38 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stabil
32. ieved Two orders at the price pO on the offer and demand side could be matched against each other Therefore a trade at price pO is done The spread rises up to 3 ticks after the trade Withanawasam 2010 offer price p trade at price p0 1 Lh new best bid Figure 6 Order Book Model 21 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability The traders apply their own trading strategies i e limit order or market order to make the orders The limit order book will decide to accept orders reject the orders or place order in different ways by the limit order book rules The orders accepted by the rules will store in the limit order book according to price and time priority algorithm The trade will be executed when a counter trader Buyer Seller offer a price that could be matched with that stored in order book This trade confirmation message will both send to traders on two sides On another hand the limit order book also has to display all information about price and depth of the orders to all relevant financial agents This information will have them to make decisions to send an order The diagram in Fig 7 shows the basic system architecture Order Book Figure 7 Basic System Design of Limit Order Book Model Do Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability The sequence diagram is a type of interaction diagram that shows the pr
33. iles and get agents numbers from one of configuration files SimulationModel It is main class of this model that initializes certain number of traders and runs traders for a maxT number of times It extends SimState class used by Mason It contains main method and use loop function to make simulation run continuously The hash map is imported to this class to call the relevant maps that set up in ModelSetting class The ArrayList is implemented for scheduling the traders model agents package ContPlayer This class simulates the trader agent s behavior that generates orders and updates thresholds FarmerPatientPlayer This class simulates that the agent places limit orders according to the limit order books model MixImpatientPlayer This class is a mix of cont and farmer impatient players They place market order It uses cont to determine when to trade an on which side And it uses a double auction book to place orders PlayerInterface It is an interface class for all agents It is a steppable Mason object that helps agents connect with ModelSetting class In this class trader s id is defined RandomNum This class generates various random variables The FarmerPatientPlayer and MixPlayer will call this class for random variables model market package TradingEnviroment This class simulates a trading environment like market It generates the make price according to the orders placed by agents BookInterface It is an interfac
34. ion pops up which allows the user to start pause and run the simulation 3 Click on the run button to start the simulation 4 If need to stop or pause the run click on the on the stop button or the pause button respectively Forder Book Simulation Lingeen Huang File 2inix About Console Displays Inspectors Model ps Order Book System MSc Financial Computing Project By Lingcen Huang Order book system is a model to simulate the performance of a order book of certain asset An order book is a list of outstanding orders manual or electronic at an exchange venue i e the London Stock Exchange LSE The book records the interest of sellers and buyers in certain financial instruments It is used to decide vehich order could be fulfilled by the trade engine free j Start Pause Stop 49 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability 5 Click display tag to view the graphics Prices plot Average volatility plot Order books 6 Double click the name of graphics to display them The separate graphic for each one will pop up 7 Click the button Save as PDF to save those graphics Chart Right Click or Control Click on ChatiorMor Options MEM Title Prices plot veneni cid Me X Label Step E a Y Label Price E LogXaxis MARIS XM se LogYaxis
35. it order order book and multi agent modeling was obtained from literature review And in the system simulation a lot of conclusions could be made If system applies the simplest order book without any rules the price of asset will has a substantial fluctuation If system applies double action book with rules agents place market orders and limit orders the stability of price will be improved significantly In conclusion this project can be considered a success according to the reasons described above 43 Lingcen Huang ra Novel Order Book Rules to Promote Trading System Price Stability 6 3 Further Work In this project we developed two kinds of models One is simplest model with simplest trader behaviors Another is complex one with order book rules In the further work the description in below could be direction for further development Combination Model In this model we can combine some aspects of the two models that have been developed successfully The combination model will collect more positive characteristic of both models It continually applies double auction mechanism The two types of traders will interact via it The patient trader s behavior remains same but the impatient trader will be modeled after cont They have heterogeneous thresholds and place market orders according to the public information that is a normally distributed random variables compares with these thresholds Some traders adjust the volatility
36. ity Figure 23 Order Book for Double Auction Book Model at D 0 3 Figure 24 Price for Double Auction Book Model at D 0 03 39 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability Figure 25 Average Volatility for Double Auction Book Model at D 0 03 Figure 26 Order Book for Double Auction Book Model at D 0 03 40 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability Figure 27 Price for Double Auction Book Model at D 0 003 Figure 28 Average Volatility for Double Auction Book Model at D 0 003 41 Lingcen Huang era Novel Order Book Rules to Promote Trading System Price Stability Order books 25 0 2251 20 0 17 5 15 0 12 5 10 0 Number of observations 75 501 25 0 0 Figure 29 Order Book for Double Auction Book Model at D 0 003 5 4 Analysis of Results In the cont book model the price movement is extremely huge When D 0 3 the maximum point can reach 1200 When D reduces to 0 03 and 0 003 the maximum price reduces to 25 But the price fluctuation is still significant it is more than 100 of initial price On another hand whatever D equals to the average volatility does not increase or decrease too much It keeps about at 0 08 0 036 and 0 015 when D equals to 0 3 0 03 and 0 003 All of those chart results show the price is totally unstable in this model even the D is very small In the do
37. led or executed Therefore the limit order book becomes highly dynamic and keeps changing throughout the trading day Table 1 shows a simple order book The order information is transparent to the traders Table 2 shows the order book after submitting a buy limit order at price 48 2 for 400 shares Table 3 shows a trader submitting a buy market order for 300 shares source Chen S 2007 Computationally intelligent agents in economics and finance Information Sciences 177 5 1153 1168 ASK BID Shares Price Price Shares 400 49 3 49 8 600 100 48 2 30 1 450 50 47 7 50 6 700 150 46 5 50 9 500 200 45 4 D2 100 Table 1 The simple order book ASK BID Shares Price Price Shares 400 49 3 49 8 600 500 48 2 50 1 450 50 47 7 50 6 700 150 46 5 50 9 500 200 45 4 51 2 100 Table 2 Order book after submit 400 shares buy limit orders at price 48 2 9 Lingcen Huang Ir Novel Order Book Rules to Promote Trading System Price Stability ASK BID Shares Price Price Share 400 49 3 49 8 300 100 48 2 50 1 450 50 47 7 50 6 700 150 46 5 50 9 500 200 45 4 91 2 100 Table 3 Order book after submit 300 shares buy market orders 2 2 Order Book Trading Rules and Strategies 2 2 1 Order Book Trading Rules The order book trading rules are applied to align with system and operation rules They refer to the trading system and the parameters that are applicable They are set according to change of segment and sector level system configuratio
38. ly at current best price The output graphics including price plot volatility plot and order book for these two scenarios are compared and analyzed The report starts at introducing the motivations of this topic and using multi agent simulation And then it defines the relevant concepts of order book The trading rules and trading strategies are also descried The report records the design and implement simulation model by each step The simulation results are analyzed carefully Finally the report ends with description of finding and conclusion i Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability Acknowledgements To Yao for the love and support To Toni and Chris for the guidance through out this project To my parents for spiritual and financial support Lingcen Huang eral Novel Order Book Rules to Promote Trading System Price Stability Content Chapter 1 Introduction sicssscscsccsssssssssscososscessssncsssssonsssnedsnsossssansosnenancsssssansessoaecsacensss 3 iM I ru ME M 3 1 1 1 Order Book MOE VALUGT socios ei cabe Ike REP SR PREX EIS Rr SUMI OSEE UR CUR RINT o E Dx e ERI cs 3 1 1 2 Agent based Modeling Motivation is s1 icsissssiostsssssassnnsivsdenseacationseversssedsaneves 3 1 2 Amend ODICCIVESSeaia ner Cerne aae Weg FUE PH EPA EVE EI ECC dp UE 5 1 3 Project Report OvervieW sssesssesssecssocssoossosesssesssesesocssoosssossssesssesssocsssossossssse
39. n in trading system The rules include how a member firm enters the trading system and its responsibility for the order management They could be member authorized connection or order routing The rules also cover the reversal of erroneous trades and contra requests When a trading system model is designed it is necessary to take into consideration the order book trading rules Table 4 shows these rules in details LSE 2011 Trades On Exchange trades 2000 automatically effected on an Exchange order book Order entry Access to the trading system and the responsibility of member firms 2100 Each order submitted shall be firm and subject only to the terms relating to benefit entitlements prevailing at the time of execution G 2101 The member firm shall at all times have sufficient order management systems procedures and controls designed to prevent the entry of erroneous orders GT 2102 A member firm should use the correct dealing 10 Lingcen Huang gig Novel Order Book Rules to Promote Trading System Price Stability capacity indicator D 2105 The Exchange reserves the right to refuse a member G firm s request Contra request G 2110 If a member firm submits an order incorrectly which GT is subsequently executed it may submit a request to contra the resultant trade s Exchange enforced cancellation of erroneous trades D 2120 The Exchange vi
40. ndix E Bibhography sisicacisdinmmcnninaiinaaatinaenaas 56 1 Lingcen Huang wt Novel Order Book Rules to Promote Trading System Price Stability List of Figure Figure 1 An agent 4 Figure 2 A Simple Order Book 8 Figure 3 Algorithm for Maslov Limit order generator 14 Figure 4 Action and Interaction between agents 15 Figure 5 Use Cases Diagram 17 Figure 6 Order Book Model 21 Figure 7 Basic System Design of Limit Order Book Model 22 Figure 8 The Sequence Diagram for Limit Order Book Simulation 23 Figure 9 The Workflow of the order book system 24 Figure 10 The Workflow of Insert a New Entry in Order Book to Buy Side 24 Figure 11 Class Diagram for Limit Order Book Model 25 Figure 12 System Package Diagram 27 Figure 13 Java Class Structure 30 Figure 14 Testing Process Diagram 32 Figure 15 Price for Cont Book Model at D 0 3 35 Figure 16 Average Volatility for Cont Book Model at D 0 3 35 Figure 17 Price for Cont Book Model at D 0 03 36 Figure 18 Average Volatility for Cont Book Model at D 0 03 36 Figure 19 Price for Cont Book Model at D 0 003 37 Figure 20 Average Volatility for Cont Book Model at D 0 003 37 Figure 21 Price for Double Auction Book Model at D 0 3 38 Figure 22 Average Volatility for Double Auction Book Model at D 0 3 38 Figure 23 Order Book for Double Auction Book Model at D 0 3 39 Figure 24 Price for Double Auction Book Model at D 0 03 39 Fig
41. o any errors Mason It includes all files from Mason Library It includes all libraries that will support the system simulation 4 2 1 Description of Java Class In this section the implementation approach and function of each java class will be described in details There are fifteen java classes in this model in total GUldisplay package Displaysupporter This class implement Steppable simulation engine It is used to communicate with other classes ArrayList and Vector are applied to present XY series Jfreechart library is called in this class to help draw the various charts SimulationModelWithUI This class is in charge for building and managing the GUI It builds the simulation interface for users and output different type graphics with using Jmf library The time series chart generator is used to generate price plot And histogram generator is used to generate order book chart The controller is to control all related parameters about output charts Index It is an html file that introduces the information about this simulation model It will display in about page when user run the system model package ModelSetting This class is responsible in connecting with step up files User can select one types of the set up files that load into model The class uses buffer reader to 28 Lingcen Huang Er Novel Order Book Rules to Promote Trading System Price Stability get parameter values from one of property text f
42. ocess operates with others and orders of operation We use it to analyze the logic of a complex procedure function or operation The diagram in Fig 8 displays the sequence of processes that operate during the model running period It describes a single time step For normal use this sequence would operate in a loop before simulation stops Schedule Counter Order Agents Trader Book Limit Order l Market Order Confirm Trading Match Orders Execution Generate Trading Reject Order Display the Strategy Confirm Trading Execution Display the Figure 8 The Sequence Diagram for Limit Order Book Simulation In order to analyze the model in more details we use the activity diagram with conjunction with sequence diagram Fig 9 provides the visual presentations of workflow of how to order book system execution Fig 10 shows how to insert a new entry in order book on buy side The sell side workflow is similar 23 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability w Is Order Under Order Book Rules Figure 9 The Workflow of the order book system Macth The Bes Sell Offer Yes Is Sell Order Book Empty Figure 10 The Workflow of Insert a New Entry in Order Book to Buy Side 24 Lingcen Huang ih Novel Order Book Rules to Promote Trading System Price Stability After analyzing the system architecture operation processes and
43. offer at the moment while trader B can buy a share of stock at price 53 best sell offer at the moment The limit order entry will be deleted from BUY SELL book when any trade happens For instance if trader A sells a share with limit order book they will get 50 and the top of BUY book 50 will be deleted from the list The last traded price will be set as 50 Trader C must place a 50 buy order to get a share stock and give 50 to trader A The current model is simplified as trading in one stock share But in the real world a limit order book should hold a large number of different shares for every book entry 8 Lingcen Huang eru Novel Order Book Rules to Promote Trading System Price Stability In the modern market most places apply an electronic double auction limit order book A trader could choose to submit a market order or a limit order Chen 2007 A market order makes sure that the execution is done immediately but cannot control its price A limit order can control the execution price but cannot guarantee when the execution could be done In the limit order book depending on market rules the limit orders can be executed in two types price priority and time priority Ask orders that has lowest price will be fulfilled firstly a sell market order gets executed at best bid price and vice versa During a whole trading day new limit orders keep adding into the order book current orders in the book are continuously cancel
44. ook Rules to Promote Trading System Price Stability Chapter 2 Background and Research Information This chapter introduces the detailed background about order book and agent based models The first section gives the order book definition limit order book rules and trading strategies by reviewing past published paper The second section discusses more information about agent based models 2 1 Order Book 2 1 1 Order Book Definition An order book is a list of outstanding orders manual or electronic at an exchange venue i e the London Stock Exchange LSE The book records the interest of sellers and buyers in certain financial instruments It is used to decide which order could be fulfilled by the trade engine The relevant definitions of terms in an order book are listed as follows Marco 2007 Price levels If several orders have the same price they could be called as a price level That means all orders on that price level have opportunities to fulfill that ask when an ask on that same price level comes Cross the book The orders could be matched according to the interest of sellers and buyers as an order book is a part of the trade matching engine When there is an order whose ask price is lower than the highest bid the order should be fulfilled immediately If it is not fulfilled and still a part of the order book due to errors this situation is referred to as crossing the book Top of book Top of book is th
45. or 2 Situated all agents live in an environment and interact with each other 3 Goal directed each agent has goal to achieve 4 Autonomous and self directed each agent has independent function 5 Flexible an agent can learn and adapt its behavior according to experience The details show in Fig 1 Environment Agent Attributes Behavioral rules Memory Resources Decision making sophistication Rules to modlify behavioral rules Figure 1 An agent We consider appling agent based modeling because the issue we study in this project is complex The system that we try to analyze and build is much complex due to the agents inter dependence A traditional modeling approach will not adequately model this system Furthermore this project will involve a large amount of data agent based modeling could provide much better approach to organize those data As a conclusion applying agent based modeling in this project is beneficial due to the reasons listed below Macal 2006 4 Lingcen Huang Rea Novel Order Book Rules to Promote Trading System Price Stability The project has decisions and behaviors that can be defined discretely The project needs agents who adapt and change their behaviors The project needs agents who learn and engage in dynamic strategic behaviors The project may require agents having a dynamic relationship with other agents and agent relationships that form and dissolve
46. platform independent Each class is created for a special purpose All of these classes can be used more times to become an object that can be then be executed The Mason software and java library are applied to output simulation results in visual charts and graphics In order to complete the project each step should be done successfully and in time Firstly we learned how to use Mason and combine it with Java Then we searched for related papers and did literature review After the background and related definitions are clear we analyzed the user requirements and functional requirements The next step was to design system based on requirements When design was completed we developed java code to implement and test system If the system could pass the testing the final step of project is to generate simulation results and analyze them 6 2 Conclusion The main goal of this project was achieved successfully at end of project To achieve this main goal every objective in each step is done during the whole project period All user requirements from UR1 to UR7 that described in chapter 3 are achieved successfully in this system All non functional requirements are achieved in this system too The functional requirements except FRIO are met in system implementation The priority of FR10 is could have that means it is optional This function can be done as a future work During the whole project process a large mount of knowledge about lim
47. rate Orders Em o iran dae acceptOrder bool cancelOrder bool p L step void A ContPlayer perg double Figure 11 Class Diagram for Limit Order Book Model 25 Lingcen Huang ra Novel Order Book Rules to Promote Trading System Price Stability Chapter 4 Implementation and Testing This chapter explains the implementation and testing process First section gives the descriptions about how to apply functionalities including initialization input and output The second section develops a way to test the system 4 1 Software Application Used Java SE JDK 6 Update 26 This is a java platform and can be run in any operation system It is selected as MASON is applied in this project Its agent modeling toolkits are developed by pure Java language NetBeans 7 0 This is a programming developer tool that both refers to an integrated development environment IDE for developing with Java and a platform framework for Java desktop application It provides the environment to write codes compile and debug them It helps improve the efficiency of implementation work MASON This is a kind of multi agent environment developed in Java It provides enough functionality for this project simulation It is uploaded into Neatbeans to help me develop many agents and output visual results With Mason the implementation work improves a lot in efficiency Creately Desktop This is a diagraming application from Cinergix Pty L
48. s executed through a trading day is divided into N smaller child orders The child orders could have different sizes Each child order is submitted as a market order at regular intervals during the trading day The time intervals are mostly set as 15 minutes or 30 minutes 12 Lingcen Huang Ier Novel Order Book Rules to Promote Trading System Price Stability The limit order strategy is a child order submitted as a limit order through a trading day Cui 2010 The trader must determine the limit price lifetime and amendment frequency for each child order A child order is set to have a T minutes lifetime At minutes amendment frequency and price the same as the best available price The limit order is placed at best price ask bid at submission time It will be amended to best price if it is not fulfilled during At minutes after submission The process will keep going in At interval till T minutes At T minutes if it still has uncompleted orders left they will be executed as market orders by crossing the bid ask spread As described above it is modest level to set the limit price as best available price Except modest level the trader also could submit the orders in aggressive level and passive level An aggressive limit order is to sell buy at the price that is one tick size below above the best ask bid price A passive limit order is to sell buy at the price that is one tick size above below the best ask bid price For a
49. sert limit order to buy if Empty L Sell price L Price RndI 1 t break else price L Sell Price RndI 1 t break InsertLOBuy L Buy price break else buy at lowest sell price LO RemoveFirst L Sell L Price LO Price break break else Sell behaviour if EmptyCL Buy orRnd 0 1 xplo insert limit order to sell if Empty L Buy price L Price RndI 1 t break else price L Buy Price RndI 1 t break InsertLOSell L Sell price break else sell at highest buy price LO RemoveFirst L Buy L Price LO Price BNRREBRESBENAREBRASEENRRAEERE S owen anne wne Figure 3 Algorithm for Maslov Limit order generator 2 3 Agent based Modeling An agent based model is effective model simulating the actions and interaction of individual agents with a view to assessing their effects on the system as a whole Before we build an agent based model we should identify agents and their behavior identify the relationships between agents and get the necessary agent related data At a general level building an agent based model should follow the five steps described below Macal 2006 14 Lingcen Huang Beau Novel Order Book Rules to Promote Trading System Price Stability Step 1 Agents Identify the types of agents and their attributes Step 2 Environments Identify the environments that the agents will stay in and interact with Step 3 Agent Methods Set up the methods by
50. stem must be defined We design this system for traders bank dealers financial agents and financial market researchers The parameters they need to specify could include the best ask price the best bid price the ask depth the bid depth and the spread Hence the system has 2 kinds of actors One is system user another is system developer The developer is in charge of designing developing calibrating and testing the system And the user is in charge to use the system functions to analyze their real order book problems They also need a summary of this system s functionality load the system and run the model A use case diagram and task description are introduced as follows Load Model ren sa i System User A Exoprt Results Run Simulation ZN lt lt implementation gt gt vu Apply Algorithms p Test Model System epe Edit Model M Create Model Save Model Figure 5 Use Cases Diagram 17 Lingcen Huang Ier Novel Order Book Rules to Promote Trading System Price Stability Load Model The user chooses a developed model from the computer The computer provides a list of models with different locations The user finds the required model and selects it The system loads the selected model and actives it The details will be shown in the user manual View Statistics The user chooses to see the output in a statistical way The model should display all statisti
51. t order book model is called cont book In this model all trades will follow the same trading rules and have the same behavior All traders are assigned volatility threshold independently In the each period the traders receive a common signal The signal is seen as public information that simulated by random variables generator To response the signal each trader decide to sell buy or sit out the period After that the market determines the excess demand and reaches the market clearing price by means of a market impact function At last traders update the threshold so that can match the absolute value of the return rate for the current period The more complex order book model is called double auction book This book organizes the limit orders and makes optimization between market order and limit order To achieve this aim the rules should be applied in order book This model has two types of traders One is to place market order another is to place limit order This model explains a large part of the price and spread diffusion of actual stocks given an order flow rate which means that the double auction structure has a huge impact on the nature of market movements When users simulate different type of models they also need change the setup file that connects to model in modelSetting class The set up files specify the order book type applied and number of traders DABook properties file is for double auction book model SBook properties file
52. tation It will create itself a sub folder mason while extracting 4 Download the file libraries zip from http cs gmu edu eclab projects mason and extract into dissertation That will create the sub folder libraries 5 Download the file jdk 6u26 nb 7 0 windows ml exe form http www oracle com technetwork java javase downloads jdk netbeans jsp 14293 1 html 6 Install JDK NetBeans Bundle into computer 7 Choose the Download Java 3D 1 5 1 Software option from http www oracle com technetwork java javase tech index jsp 138252 html and download the file java3d 1 5 1 windows 1586 exe 8 Install the Java3D framework into computer 9 From your C Java folder copy the sub folder Java3D into dissertation 10 Copy the JAR files itext 1 2 jcommon 1 0 0 jfreechart 1 0 1 and jmf from D dissertation libraries to D dissertation mason Copy the JAR files j3dcore j3dutil and vecmath from _ D dissertation Java3D 1 5 1 lib ext to D dissertation mason 45 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability Appendix B User Manual Create Project and Import Codes 1 In the main window select File gt New Project Then choose the category Java and the type of project Java Application Click Next Steps 1 Choose Project 25 Bue Maven NetBeans Modules Java Project with Existing Sources a Java Free Form Project
53. td to help me create use case diagram sequence diagram activity diagram class diagram Gantt chart and anything visual needed for the project Java Libraries Java library is the compiled byte codes of source code developed by the JRE implementer to support application development in Java The libraries used for developing the system are as follows 26 Lingcen Huang Novel Order Book Rules to Promote Trading System Price Stability 1 Ltext It is a library allowing the user to generate and manipulate PDF documents 2 Jfreechart It is a free java chart library that helps users to display professional quality charts in their simulations 3 Jcommon It is a java class library used by Jfreechart It is used to support user interface classes for displaying information about applications and serialization utilities 4 Jmf It is a java library called java media framework The developer uses it to add audio video and other time based media to java applications and applets 5 Vecmacth it provides a powerful way to handle and manipulate coordinates and vectors being compatible with how the Source engine handles them 4 2 System Implementation The implementation of system was developed with Java Java libraries and Mason software It is a complex system that involves the object orient design The Mason software is applied to output visual results Five packages are created according to system structure The model p
54. tem Requirements There are two different types of system requirement one is functional requirement another is non functional requirement They are distinguished so that each refers to the implementation of one user requirement and the technical design of the system The functional requirements define a function of a system or its components while the non functional requirements describe criteria used to access an operation of system rather than specific behaviors They could often be the limits on the system that are qualitative measurements such as cost quality and performance The functional non functional system requirements are listed in tables 5 and 6 below The priority of functions can be defined into three levels M Must Have crucial to the system S Should have and C Could have optional is also given for each requirement ID Functional Requirement Priority 1 The model shall implement multiple agents at least two agents M order book and trader 2 The model shall implement an order book for the market and limit M orders are placed and matched 3 The model shall implement order book rules to limit the orders M 4 The model shall list bid ask price from high to low from low to high in M separate columns 5 The model shall use the visualization built into MASON to show the M model during a run and after 6 The model shall represent the ask price in red line
55. uble auction book model the simulation results are totally different However the value of D is large or small the price fluctuation is extremely small The lager the value of D is the closer to initial price the average price is Whatever the value of D is the average volatility is in a very small value The order book charts are displayed as expected They are very similar in three different values of D They all looks like sinusoidal distribution Most buy and sell orders are placed in middle of prize axis Only a few of sell orders 1s at a very high price while also a few of buy orders is placed at a low price This order distribution from charts also means the price of asset should be stable All in all we can make conclusion from the charts from above A limit order book with rules can help improve the price stability significantly 42 Lingcen Huang Lord Novel Order Book Rules to Promote Trading System Price Stability Chapter 6 Conclusions and Future Work This chapter summarizes the findings from chapter 5 and gives final conclusion The future work is discussed in the second section of the chapter 6 1 Project Summary The main aim of project is to develop an order book system that can place market and limit order then match the orders A set of rules must be implement into order book system to increase the market price stability Java language is used to develop codes due to it is object oriented programming language
56. ure 25 Average Volatility for Double Auction Book Model at D 0 03 40 Figure 26 Order Book for Double Auction Book Model at D 0 03 40 Figure 27 Price for Double Auction Book Model at D 0 003 41 Figure 28 Average Volatility for Double Auction Book Model at D 0 003 41 Figure 29 Order Book for Double Auction Book Model at D 0 003 42 List of Table Table 1 The simple order book indicate the source of the table 9 Table 2 Order book after submit 400 shares buy limit orders at price 48 2 9 Table 3 Order book after submit 300 shares buy market orders 10 Table 4 Order Book Trading Rules of London Stock Exchange source London Stock Exchange LSE 2011 Rules of the London Stock Exchange Effectivel 35 43 _11 Table 5 Functional System Requirements 20 Table 6 Non Functional System Requirements 21 Table 7 Explanations of Parameters 31 Table 8 Scenario 1 Parameters Setting 34 Table 9 Scenario 2 Parameters Setting 34 2 Lingcen Huang ral Novel Order Book Rules to Promote Trading System Price Stability Chapter 1 Introduction This chapter gives a brief introduction to the project It explains the motivation for the project and its objectives The motivation relates to financial concepts and multi agent models 1 1 Motivation 1 1 1 Order Book Motivation In the modern time many of the important stock exchanges like London Stock Exchange New York Stock Exchange and Tokyo Stock Exchange depend partially on limit order
57. which agents attributes are renewed to make response to the agent interactions with environment or interactions between agents Step 4 Agent Interactions Use the methods to make interactions under control The related issues can include which agents interact when they interact and how they interact during the simulation Step 5 Implementation Implement the agent based model into the computational software As we want to simulate an order book system and study its behaviors an agent based model should be built The order book BUY trader and SELL trader are three individual agents Fig 4 shows their actions and the interaction between them Trade Conformation Trade Conformation Information Information Figure 4 Action and Interaction between agents 15 Lingcen Huang ra Novel Order Book Rules to Promote Trading System Price Stability Chapter 3 Order Book System Design This chapter shows designs of limit order book system process in details They include gathering requirements use case analysis and limit order book design The class diagram activity diagram sequence diagram are applied to show the design process 3 1 Guidance from Related Research At the moment a large amounts of trading platforms simulations are developed using agent based modeling most utilizing libraries Researchers and research students design many of them One of the recent most helpful simulators is designed by Preis
58. workflow the next step for design system is to identify classes We will combine responsibility design approach and data driven design approach to build a full class diagram Fig 10 This diagram will describe the structure of the order book system by showing the all classes their attributes methods and their relationship among the classes LimitOrder assetlD int type DisplaySurpporter SimulationModelWithUI reae double E occ acfChart entryTime load void acfFrame inital void orderHist orderHistFrame priceChart priceFrame returnChart nextUpdate ContBook getSimulationInspectedObject object load void Paa vor DoubleAuctionBook quit void setUp void start void ModelSetting Booklnterface use Creat and Setup Simulation getBidPrice double SimulationModel getAskPrice double runiD getAvarageTradePrice double paremeterMap getBalance double getBuyOrder double start void getSellOrder double cleanUp viod getSpead double getVolume double executeMarketOrder double placeLimitOrder bool cancelLimitOrder bool setID void lt lt use gt gt lt lt use gt gt cleanUp viod expireTime getStatus String LimirOrder quantityCheck bool quit void FarmerPatientPlayer Price Execution Feedback TradingEnviroment MyWorld OrderBook MixPlayer Playerinterface Gene
59. ystem behavior should be visualized so that the results could be shown directly 1 3 Project Report Overview Chapter 1 gives a brief introduction to the project It explains the project s motivation and objectives The motivation relates to financial concepts and multi agent models Chapter 2 introduces the detailed background about order book and agent based model The first section is to give the order book definition limit order book rules and trading strategies by reviewing past published papers The second section is to discuss more information about agent based models Chapter 3 presents the design of the limit order book system process in details These include gathering requirements use case analysis and limit order book design The class diagram activity diagram sequence diagram are applied to show design process Chapter 4 explains the implementation and testing process First section gives the descriptions about how to apply functionalities including initialization input and output The second section develops a way to test the system Chapter 5 analyzes the experimental results This chapter displays all resulting output from the system These results are compared and analyzed The findings are also discussed in this chapter Chapter 6 summarizes the findings from chapter 5 and gives final conclusion The future work is discussed in the second section of the chapter 6 Lingcen Huang eral Novel Order B

Download Pdf Manuals

Related Search

Related Contents

"取扱説明書"  RayBio Biotin Label-based Human Obesity Antibody Array 1  user manual    8461 Printer - Mettler Toledo  

Copyright © All rights reserved. Failed to retrieve file