An improved procedure for simulation of demand-responsive transport services in agent-based transport simulation framework

Abstract

Demand-responsive transport (DRT) is gaining popularity in transport planning. While more efforts are devoted to the optimization of DRT operational strategies, the challenges of simulating DRT in the context of transport planning, such as long simulation time and the high sensitivity of DRT systems to demands, receive relatively less attention. In this study, we develop an improved procedure for simulating DRT in agent-based transport simulation frameworks, based on the concept of estimation and teleportation. A DRT estimator is constructed based on real-world data and the knowledge of the transport system. During the simulation, the estimator provides estimated trip data for each DRT ride. The agents are then teleported to the destination based on the estimation. By removing the explicit modelling of DRT during the transport simulation, many challenges associated with conventional simulation approaches can be addressed. This approach can also be applied to policy studies of other emerging transport modes.