Conflict Management Of Rail Resources In Vehicle-centralized System Based On Event And Cooperative Game

The vehicle-centralized train control system,which is oriented by technologies such as system simplification and train intelligence,has been focused day by day because of the advanced construction of western railway network.The demand for engineering application is also particularly urgent.In recent years,the resource conflict problem based on decentralized routing control mode of this new scheme has become a key issue in related researches because of its restriction on operation efficiency.Based on the onboard interlocking system,train is designed to accomplish its own routing work in new train control scheme mentioned above.This decentralized management structure about limited rail resources fits with game theory’s design philosophy that each participant has the ability to make independent decision,opportunely.This paper sets up a microscopic operation event model based on directed topology graph theory firstly,describes the evolution of train spatio-temporal requirement about operating resources,proposes the criterion of conflict prediction,then devises conflict resolution algorithm by combining the cooperative game theory with dispatching principles.The conflict-free reordering solution is searched iteratively.The main contents of this paper are summarized as follows:(1)The routing procedure of onboard interlocking system is devided into various stages with the granularity of rail resources.Mechanism of resource conflict based on vehicle-centralized train control design is also analyzed detaily.Furthermore,the researching scheme are summarized based on features of conflict problem.(2)On the basis of directed topology graph theory,the usage of rail resources is described formally.Then,the operation event model representing the evolution of train’s spatio-temporal requirement about resources is also constructed.Finally,the temporal topology theory is used to check the overlap between demanding events about same rail resource.The criterion of conflict prediction is proposed.(3)Combining the conflict dispatching principles and spatio-temporal information mentioned above,a resolution algorithm based on cooperative game theory is designed.The operation performance of relevant trains under various priority decisions combinations is described as gaming profit.The resolution is transformed into optimization problem with the goal of maximizing the overall operation profits of regional trains.The resolution of resource requisition priority is searched.(4)The simulation test cases are developed based on Matlab environment operation data of large-scale junction station in Changsha.By configuring conflict scenarios with various operational densities and disturbance durations,the decision strategies of conflict prediction and resolution algorithm proposed is analyzed.Furthermore,indicators including 1)the maximum delay;2)the average delay;3)the scale of influenced trains;and 4)computing cost are adopted to evaluate trains overall operations in test cases.The performance of model proposed is also conmpared with traditional conflict resolution strategy and typical dynamic programming algorithm.Focusing on conflict scenarios of the most complex test case group within 60 minutes,the maximum and average delay according to the resolution of algorithm proposed are820 s and 192.3s,which are reduced by 32.1% and 13.3%,respectively,when compared with traditional strategy.The superiority of algorithm is obvious.When compared with dynamic programming’s performance,though the maximum delay of algorithm proposed is 7.1% higher,the average delay is reduced by 4.9%,the computing time cost is 15% lower,and the scale of influenced trains is also reduced by 23%.Generally speaking,the algorithm proposed in this paper exhibits a good trade-off between acceptable computational cost and disturbance suppression effectiveness.Its feasibility is verified comprehensively.This paper has 40 figures,12 tables and 116 references.