Two-Step Optimization of Urban Rail Transit Marshalling and Real-Time Station Control at a Comprehensive Transportation Hub

Abstract Urban rail transit connecting with a comprehensive transportation hub should meet passenger demands not only within the urban area, but also from outer areas through high-speed railways or planes, which leads to different characteristics of passenger demands. This paper discusses two strategies to deal with these complex passenger demands from two aspects: transit train formation and real-time holding control. First, we establish a model to optimize the multi-marshalling problem by minimizing the trains’ vacant capacities to cope with the fluctuation of demand in different periods. Then, we establish another model to control the multi-marshalling trains in real time to minimize the passengers’ total waiting time. A genetic algorithm (GA) is designed to solve the integrated two-step model of optimizing the number, timetable and real-time holding control of the multi-marshalling trains. The numerical results show that the combined two-step model of multi-marshalling operation and holding control at stations can better deal with the demand fluctuation of urban rail transit connecting with the comprehensive transportation hub. This method can efficiently reduce the number of passengers detained at the hub station as well as the waiting time without increasing the passengers’ on-train time even with highly fluctuating passenger flow.