Research On Optimization Plan Of High-speed Railway Train Based On Dynamic Passenger Flow Distribution

The formulation of high-speed railway passenger is an important part of high-speed railway transportation and a key factor in determining the quality of high-speed railway train operations.By formulation a reasonable high-speed railway train operation plan,it is conducive to improving the service quality of passenger transportation and enhancing the competitiveness of high-speed railway in the passenger transportation market.This thesis studies from the following aspects.Analyze the high-speed train operation plan,and introduce the main contents,including the train type,number of trains,train marshaling and stop planning.The influencing factors of the high-speed train operation plan are summarized.Through the analysis of the impact of passenger flow dynamic demand,refined the basic principles and formulation process of the train development plan.Due to the dynamic demand of passenger travel,the attributes of the data are largely reflected with strong timing dependencies.Therefore,this thesis establishes a prediction model for high-speed railway passenger flow based on LSTM(Long-Short Term Memory)deep learning network.The comparison of experiments also proves that LSTM can predict the passenger travel dynamics well in the prediction of railway passenger flow,and has high precision,which lays a foundation for the optimization of train operation plan.Aiming at the problem of train operation in the train operation plan,a passenger flow allocation model based on the psychological change of passenger ticket purchase process is proposed.Secondly,in order to realize the combination optimization of the number of trains,the type of trains,train marshaling and stop planning,according to the nature of the decision variables,this dissertation proposes a thought,which is the main decision-making variable of the type of train,the number of trains and the scheme of the train.The number of trains is used as an auxiliary decision variable.A multi-objective optimization model was established to maximize the economic benefits of the railway sector and maximize the market effect.In order to obtain better optimization results,this thesis proposes a multi-population adaptive multi-objective differential evolution(MA-MODE),and designs an adaptive parameter control strategy for the different degrees of individual evolution in the population.The differential evolution algorithm can be applied to the optimization of 0-1 variables.At the same time,a selection operation method for solving multi-objective problems is designed.The efficiency of algorithm optimization is verified by the test function and the example of high-speed railway train scheme optimization.It is worth mentioning that this thesis also designs a heuristic algorithm based on OD passenger flow,site weight and the characteristics of two types of high-speed trains to obtain a better initial solution,thereby further improving the computational efficiency and Ability to find excellence.