Multi-group Parallel Multi-agent Cooperative Energy-saving Optimization Algorithm For High Speed Train

China's high-speed railway has achieved great development in recent years.The passenger dedicated line has extended continuously and the number of trains has been growing.Besides,China's latest high-speed bullet trains named "Fuxing CR400AF"and "Fuxing CR400BF" have gone into service with 350 kilometers per hour traveling between Beijing and Shanghai.Under the background of the rapid development of high-speed railway,the optimization of high speed train energy saving operation has become one of the important issues to be solved in the construction of intensive society.At present,the research on energy-saving optimization of trains mainly concentrates on single-train,ignoring the interaction between trains,and does not consider the elastic adjustment strategy among trains.At the same time,the optimization methods of train energy-saving mainly focus on the construction of cost function model and the application of intelligent algorithm.The research on multi-mode parallel multi-agent cooperative optimization algorithm for train group is relatively scarce.This thesis fully considers the actual demand of high-speed train with small interval and high density and studies the multi-group parallel multi-agent cooperative optimization method.The proposed approach is analyzed based on the line characteristics of high-speed trains to access the energy saving and punctuality of train group.The main work of this thesis is as follows:(1)In the process of dynamic running of high speed train,this thesis further studies the train dynamics theory,and a hybrid dynamic model of high-speed train in moving block system is established.Based on self-perception and information sharing of train operation state,a multi-agent information interaction model is constructed.And a resilience evaluation model of high-speed train group is proposed,which provides effective support for subsequent research.(2)Focused on the problem of train group offline operation optimization,the thesis takes minimum energy consumption of high-speed train group as the research objective based on the optimal energy-saving control strategy between railway sections,and a multi-objective optimization model of train group energy-saving operation is established.Multi-group parallel multi-agent seeker optimization algorithm and differential evolution are used to solve the model,which realizes the train group offline cooperative optimization.The performance of the two algorithms is compared and analyzed.The simulation results show that the multi-group parallel multi-agent differential evolution algorithm can solve the offline optimization problem more effectively.(3)Aiming at the uncertainty of train running process caused by the complex railway environment and the constraints of track distance between trains,the resilience evaluation model is established to realize the real-time assessment of train running deviation.At the same time,the assessment results are used as the trigger condition for the online cooperative optimization algorithm of train group.According to the conversion of train operation conditions,the online planning of the operation strategy of the remaining train group is realized,which ensures the safety and punctuality of train operation and improves the overall energy efficiency of train group.Finally,this thesis simulates and verifies the algorithm based on the real line data of the "Wuhan-Changsha South" section of Wuhan High-speed railway.The results indicate that the proposed multi-group parallel multi-agent optimization algorithm can realize the offline cooperative optimization and online cooperative adjustment of train group,which ensures the safety and punctuality of train group.Compared with the results that have not been adjusted,the online cooperative adjustment strategy could save 6.43%of the whole running energy of train group.The whole time deviation of train group is reduced from 63 s to 4s,therefore the validity of the algorithm is proved.