Train Optimized Control Based On Genetic Algorithm And Fuzzy Expert System

With the development of high speed railway in our country, train speed and transportation density are increasing continuously. Traditional manual train operation methods are out of requirement. At the same time, train assistant operation system which makes great use of advanced computer technology can get reasonable solutions by simulating and forecasting the process of train running. The solutions not only assure the safety of train and the comfort of passengers, but also increase train speed and save energy. The research about train assistant operation system is powerful support to current manual train operation. And it is the necessary requirement of railway automation in my country.This paper proposes an on-line fuzzy expert system and a genetic algorithm (GA) for train optimized operation, through study of some basic train operation strategies and evaluation functions, which have good performance in practice. The main contents of the thesis are as follows:(1)Train operation strategies are expatiated and analyzed from three aspects, namely energy saving, punctuality and comfort, based on predecessor's work. Some unique ideas are suggested for complicated train operation.(2) A solution scheme of fuzzy expert system is presented based on analysis of train movement. Expert knowledge database and rules database are established from experiences of excellent train drivers with decision of exact and fuzzy variables. And a fuzzy expert system mechanism was built through discrete even burst mode.(3).The model of train operation problem and the new conception of "Key Points" are given in this paper. The control strategies of train control in inter stations are described by a table of "Key Points". Then the table of "Key Points" is encoded as the beginning procedure in whole GA implement. The results of GA optimization are satisfactory for energy saving, punctuality and comfort by comparisons between original and optimized train operation graphs.(4) Combined with the on-line fuzzy expert system, a train assistant operation system is designed and realized in this paper. And It could be operated on TrainOperation Simulator.The research is the basis of Automatic Train Operation (ATO) and means great to realize full functional ATO in our country.