Study On Multi-block Train Plan With Genetic Algorithm Of Railway Station

A better train plan has advantages of abating the scarcity of transport capacity, speeding up the wagon turnaround, reducing transport costs. Optimizing the train plan be propitious to raise the level of railway organization which also has great influence on transport efficiency and service level.This paper mainly discusses the train plan problem. Firstly, it analyzes the research situation of the train plan based on the background of the problem. Then, train planning of single trains and multi-block trains are introduced. As an important mode of railcar operation, multi-block trains have the advantage of reducing accumulation time, cutting down shunting on the way and speeding car movement which remind us to consider the single trains and multi-block trains at the same time when train planning at technical service stations and a integrative model of train plan is built to maximize the technical benefit. With the increasing of the service stations, computational complexity of the integrative train plan will increase faster after analyzing this model. Because of this, heuristic algorithm has to be used to solve this problem. Then, considering the genetic algorithm (GA) is simple, robust, rely less on initial solution, global optimizing capability, GA with single parents, GA with two parents, adaptive GA with two parents are achieved in the platform of Visual C++6.0 that are designed for train plan of single trains and integrative train plan. At last, search performance of this three kinds of GA are compared, so do technical benefit of this two kind of train plan.After researching this problem, the author gets that adaptive GA with two parents and random disturbance mechanism performs good on this problem which can get the satisfy answer. Besides, adaptive train plan gets more technical benefit than the train plan of single trains.