Research On Strategies And Methods For Train Intelligent Rescheduling In Complex China High-speed Railway Network

High-speed railway plays a key role in the national comprehensive transportation system.It has been developed rapidly and vigorously in recent years.The high-speed centralized traffic control is the brain and nerve of the high-speed railway transportation.The scale of the high-speed railway network is expanding and the structure of the high-speed railway network is more complex.When the traffic is disturbed by some uncertain factors,it is a great challenge for CTC system to make the trains run in order,minimize the totals delays and the impact of delay and improve the satisfaction of passengers in time as soon as possible.The research on strategies and methods for train intelligent rescheduling in complex China high-speed railway network is conducive to develop the level of automation and intelligence,keep railway operation safety and improve the efficiency of traffic transportation and operation.It is the inevitable trend and urgent demand for the intelligent development of China high-speed railway.According to the insightful knowledge of China train dispatching,the actual demand of train rescheduling and the studies of correlational fields at home and broad,the strategies and methods for train intelligent rescheduling in complex China high-speed railway network is researched in this paper.First,China high-speed railway network is researched based on complex network theory.Its topological structure diagram and statistical characteristic parameters are obtained by the software of Pajek.The topological structure,characteristic attributes and the characteristic attributes based on trains from the multiple perspectives are studied.The SIS delay propagation model is set up by the theory of complex network to research the behavior of delay propagation in the complex China high-speed railway network.Secondly,the minutes that the train running in section is predicted based on the machine learning algorithm.The factors that delay occurs and be decided are analyzed and summarized from the perspective of train dispatching diagram.The huge mass of space-time train running data in CTC system is mined effectively based on machine learning algorithm to predict the minutes that the train running in section.The prediction model based on random forest algorithm is set up in the condition that this section has sufficient historical train running data.The prediction model based on transfer learning algorithm is set up in the condition that this section has insufficient historical train running data.It could realize the prediction crossing different train running sections or different train dispatching section.Thirdly,the strategies and methods for train rescheduling in a single train dispatching section are studied.The characteristic attributes of train dispatching section based on trains are analyzed from multiple perspectives.The viewpoint that different rescheduling strategies and objectives should be adopted in different train dispatching section is proposed.The different train dispatching section is the train dispatching section with different location,different characteristic attributes based on trains or passengers,different interference incidents,different management or evaluation methods and so on.The common interference incidents and relevant rescheduling strategies are explained detailedly.The case base is set up by the cases that extracted from the huge mass train running date in CTC system.The train rescheduling strategy is selected by historical experience from the case base model of train rescheduling strategies.The train rescheduling model for a single train dispatching section is set up.The total objective is designed by the characteristic attributes of trains and satisfaction of passengers in overall consideration.According to characteristics of dispatcher's decision,the decision method based on prospect theory and random intuition fuzzy is adopted to choose the relevant weights of total objective.The chaotic firefly algorithm based on self-logical mapping and mutative scale is adopted to solve the train rescheduling model.Finally,the model and method of train rescheduling between train dispatching sections in railway network are studied.This problem is studied based on the theory of large scale system.The hierarchical model of train rescheduling in railway network and the synergetic model of train rescheduling between train dispatching sections are set up with the guiding idea of decomposition and synergy in the theory of large scale system.The improved max-min ant colony algorithm is optimized by four strategies.And it is adopted to solve the model.Combining with the studies in front,the goal that train intelligent rescheduling in complex China high-speed railway network has been realized.