Multiple plan approach for a dynamic dial-a-ride problem

In densely populated areas, ridepooling services are becoming more and more popular and can contribute to promote sustainable mobility concepts. In this paper, we study a dynamic dial-a-ride problem and propose a novel two-phase multiple plan approach. By maintaining multiple plans simultaneously over the course of the optimization procedure, we increase the likelihood and speed of a successful insertion for dynamically arising customer requests. Additionally, we present a new metric to replace the commonly used distance minimization as guidance within the optimization procedure. In order to examine the impact of certain instance characteristics on the quality of the solutions, we evaluate our design decisions on various instance sets. The results show that the proposed method consistently outperforms commonly used single plan approaches.