Research On The Newly-built High-speed Railway Train Operation Plan Based On Column Generation Algorithms

Railway transportation corridor is an important form of China's transportation system.With the high-quality development of economy and the gradual operation of newly-built high-speed railway passenger dedicated line,the railway transportation corridor in some areas will gradually form a new pattern of "six-line separation of passengers and goods",which is the expectation and vision of the future development of the railway system in China's Medium and Long-term Railway Network Planning.This new transportation corridor pattern will generate many problems which are needed to be solved such as how to scientifically predict the passenger flow and how to compile the train operation plan for the newly-built high-speed railway.The accuracy of passenger flow forecasting and the rationality of train planning will directly affect the whole railway system and the optimal allocation of transportation resources for a transportation corridor.Therefore,passenger transportation is taken as the research object in this paper,and passenger flow forecast train operation plan of new high-speed railway is taken as the starting point and end point,which provides theoretical support and practical guidance for solving the increasingly frequent problem.The main research contents are as follows:(1)Firstly,sort out and analyze the key concepts and clarify their meanings in the paper,and systematically analyses the evolution mechanism of passenger transport supply and demand and passenger travel behavior,which provide a theoretical basis for passenger flow prediction in the future.The preparation methods and principles of the train operation plan are introduced in this paper,which provide ideas for the construction of the operation plan model.(2)Secondly,in order to accurately predict the distribution of passenger flow in the newly-built high-speed railway,comprehensive transportation corridor is taken as the research object from the point of view of comprehensive transportation.A four-stage forecasting method of "passenger flow generation,passenger flow distribution,passenger flow induced,and passenger flow sharing" is constructed in the paper.The combination forecasting model,improved gravity model,induced traffic model and multiple Logit models are used respectively.In order to calibrate the parameters and solve them in the model,the Bayesian network,grey prediction,least square method and maximum likelihood estimation algorithms are designed by using MATLAB.A set of relatively scientific and reasonable prediction framework of passenger flow time and spatial distribution is formed.The predicted passenger flow of high-speed rail provides the data basis for the subsequent train operation plan,and the direct matching between passenger flow and train flow is realized.(3)Then,in order to compile the passenger train plan for the new high-speed railway,a two-stage design model of the high-speed train operation plan based on the column generation algorithm is constructed.In the first stage,the train alternate set generation model is established.The objective function is to minimize the total train operation cost.By constructing the sub-problem of test number,the train alternate set generation algorithm is used to solve the model.The second stage is the passenger flow allocation model.The objective function is to maximize the operating efficiency of high-speed railway passenger dedicated line,and the constraints are designed based on the principle of improving the market competitiveness of high-speed railway.Finally,the algorithm design of the new high-speed train operation plan is realized by the mixed programming of Python and Cplex.(4)Finally,taking Jiaoji Corridor and Jiqing High-speed Railway as the research objects,based on the current research and data collection,the passenger flow distribution of Jiaoji Passage is predicted and the intra-line train operation plan of Jiqing High-speed Railway is completed by using the research framework and model algorithm constructed in this paper.It proves the practicability and scientificity of the research ideas and algorithms in this paper.