Optimizing Integrated All-Stop, Express, and Short-Turn Services on a Metro Line with Capacity Constraint Considering Passenger Spatio-Temporal Path Choice

Abstract

Optimization of integrated train service is quite challenging under the situation of express trains overtaking local trains, because passenger path choice becomes more complicated and is largely unknown. To fill this gap, this study proposes a passenger spatio-temporal path choice model to replicate passenger behaviors given the real-time operation information in the metro system. Based on which, a non-linear programming model is developed to optimize integrated train services (i.e., all-stop, express, and short-turn) and the associated train timetables to minimize the total cost for a congested metro line, subject to a set of constraints (i.e., capacity, headway, service pattern, conditions of overtaking, etc.). The genetic algorithm with floating-point number coding is used to perform the optimization. It is applied to a real-world metro line in Chengdu, China, to assess its feasibility and effectiveness. The experimental results indicate that the integrated service is substantially improved in relation to total cost. The sensitivity analysis is conducted to explore the impacts of model parameters on various operation strategies. The proposed modeling approach is beneficial for transit agencies to improve their service planning.