Passenger-Induced Delay Propagation: Agent-Based Simulation of Passengers in Rail Networks

Abstract

Current work on delay management in railway networks has – to the best of our knowledge – largely ignored the impact of passengers’ behavior on train delays. This paper describes ongoing work aiming to explore this topic. We propose a hybrid agent-based architecture combining a macroscopic railway network simulation with a microscopic simulation of passengers in stations based on the LightJason agent platform. Using an initial instantiation of the architecture, we model a simple platform changing scenario and explore how departure delays of trains are influenced by delays of incoming trains, and by numbers and heterogeneity of passengers. Our results support the hypothesis that passengers’ behavior in fact has a significant effect on delays of departing trains, i.e., that passengers’ behavior in stations must not be neglected. We recommend to include these effects in up-to-date models of delay management.