Time-Dependent Dial-a-Ride Problems: Formulation Development and Numerical Experiments

Abstract

In order to satisfy the transportation demand for elderly and disabled people, providing efficient dial-a-ride systems is an important issue. Appropriate vehicle routes for dial-a-ride systems can increase transportation efficiency and reduce the operation cost. Most literatures focus on improving the vehicle routing and scheduling and assume that the travel time is a constant value. However, the travel time can fluctuate dramatically due to traffic congestion or incident. Using the constant travel time to solve dial-a-ride problems (DARP) cannot obtain the efficient vehicle routes. To consider the variation of the travel time, this paper formulates a mixed integer programming formulation for the time-dependent DARP with time window. In the formulation, the travel time variations are treated as step functions. The simulation-assignment model, DynaTAIWAN is applied to generate time-dependent travel time matrices and simulate traffic flows in the real network. Numerical experiments are conducted in a Kaohsiung network.