Object-oriented Genetic Algorithm And Its Application In Railway Traffic Control

With the developments of Railway Intelligent Transportation System (RlTS),studying an intelligent train traffic control system Which is adaptable to ChinaRailway Intelligent Transportation System is becoming more important and urgent.Now train traffic control in China is still made manually by dispatchers, bothefficiency and quality are quite low and doesn't adapt to construction of Chinarailway intelligent transportaion system. The traditional methods for train trafficcontrol, such as Operations Research and Artificial Intelligence, have their ownlimitations. In addition, multiple categories of train run on the same line in China andthis makes railway operation in China more complex than in other countries.Therefore, it is necessary to research a train traffic control method for China railway.This thesis makes improvements on coding method, genetic strategies and the way tosolve restrains for existing Genetic Algorithms. Based on this, a new method, Object-Oriented Genetic Algorithms (OOGA), is put forward and its application in traintraffic control is discussed. This paper advances an intelligent simulation method,Which combines object-oriented modeling and rule-oriented reasoning technology. Inthe end, a train traffic control simulation is carried out for Guangzhou-Shenzhenrailway and the simulation results show that the train traffic control method based onObject-oriented Genetic Algorithms is feasible.Chapter l introduces the background of this thesis and the present situation andlimitations of existing methods. The rest consists of three parts. Part l, Which includeschapter 2 and chapter 3, explains the application background and the key techniquesof the OOGA. Part 2 includes chapter 4 and chapter 5, elaborates a method forintelligent train traffic control based on OOGA in detail. Part 3 includes from chapter6 to chapter 9, presents an intelligent simulation method and demonstrates asimulation experiment to evaluate the effectiveness of the train traffic control methodbased on OOGA.The main contents are as follows:l,An Object-Oriented Genetic Algorithms that adopt object-oriented coding methodand genetic strategies is put forward at first time and the method for solvingconstraints based on expert systems, which endows GA the ability to solveexperiment constraints and extend the types of constrains that GA can deal with, isput forward, too. The OOGA is hierarchical, the top level optimizes the majorproperties and the bottom level solves the minor properties. The objcct-orientedprogramming technology is applied in implementation of the OOGA and theprogram is organized in form of objects and methods, so the inheritability andusability of program are improved.2. A new method for train traffic control based on OOGA is brought forward at firsttime, and a detail model combining both hard constrains and soft constraints fortrain traffic control on double-track lines is built and the specific ways to applyOOGA in train traffic control is studied in detail.3. An intelligent simulation method based on object-oriented modeling and rule-based reasoning technology is advanced in order to avoid the draWacks ofexisting simulation methods for train traffic control.4. The intelligent simulation environment is built for Guangzhou-Shehahen railWaylines and the effectiveness of the intelligent train traffic colltrOl method based onOOGA is validated under this environment.5. The simulation based on intelligent simulation and intelligent train traffic contrOlmethod is carried out for Guangzhou-Shenchen railWay using actllal operationdha in the year of 200l. The results show that:> The ithelligent simulation method based on object-oriented modeling andrule-based reasoning technology is feasible and effective.> The intelligent train traffic control method based on OOGA is effective interm of real-time and optimianion and it can genefatC a satisfyingdispatching decision even in...