A bi-level subway transfer simulation model based on AUTS

In urban transit management, simulation models have been widely used to evaluate the design of subway transfer stations and the passenger flow organization measures. However, most models simulate transfer stations in an isolated way, which may lead to unrealistic results, since both supply and demand of the transfer stations are highly dependent on other parts of the urban transit system. In this paper, we propose a bi-level transfer simulation model, which includes Artificial Urban Transit System (AUTS) and transfer passenger flow simulation. By simulating the whole urban transit network with explicit considerations of passenger´s behaviors on route and departure time choice, AUTS can provide dynamic passenger arrivals as input to the transfer passenger flow simulation, which then simulates the passengers´ movement inside transfer stations, and evaluates the functionality and performance of transfer stations. This bi-level model can easily captures the changes caused by other parts of the urban transit network. We develop a demonstration system which includes part of Beijing transit network and Dongdan transfer station. Computational experiment results show the advantages of the proposed bi-level model and provide valuable insights to transfer passenger flow management.