Research On An Optimization Model Of Slack Time Layout Considering Transfer In High-speed Train Timetables

Along with the construction and operation of high-speed railways, passengers expect more for the LOS of high-speed. Timeliness, punctuality and convenience of transfer have become three important standards to measure LOS of trains. Considering convenience of transfer at high-speed railway hub stations, this paper will research on timetable robust on the condition of transfer. Main work of this paper can be summarized as follows.(1) A framework of relations among transfer trains in different lines is built. Then we analyze the influence of train interference on transfer and the mechanism to improve the stability of train diagram by setting the transfer slack time on the condition of transfer.(2) We create a stochastic expected value model of slack time layout optimization considering the link and transfer based on the existing model which is the stochastic expected value model. It consists of two planning layer models, the lower layer model and the upper layer model. These two models nest inside each other and associate with each other through the slack time layout scheme and total delays of passengers. The upper layer model optimizes the slack time overall layout, and passes the optimization layout scheme to the lower layer model. Then the lower layer model adjusts the simulated diagram on the basis of the given layout scheme. We put total delays of passengers as the objective function, then measure on-line performance of initial train timetables in different scenarios through retention decisions of transfer and relief process of conflicts. And the objective function value will be fed back to the upper layer model as quality standards of the slack time layout plans. We can complete the whole layout optimization of the slack time on the condition of transfers through the coordination of the two layer models.(3) We fit for the probability density function of the train running time in sections and stop delays at stations based on the data processing and analysis of arrival and departure time in a high-speed railway line. Then we get the fitting parameter values of the double parameters negative exponential distribution for the delays, and test the fitting results. And we get different perturbation schemes we need through generate model of Monte Carlo stochastic sampling method.(4) Considering the characteristics of upper and lower layer models, we respectively select genetic algorithm method and simulation artificial method. The latter one is based on the priority to solve the bi-level programming. As to the retaining decision making process given by the lower layer model, we embed a method of delay expectations to adjust decisions on the basis of simulation.(5) We optimize the slack time overall layout on the condition of transfer in a high-speed railway section, test the convergence performance and the optimization process of the algorithm, and then give a optimization scheme of the slack time layout.