Paired cooperative reoptimization strategy for the vehicle routing problem with stochastic demands

Abstract

In this paper, we develop a paired cooperative reoptimization (PCR) strategy to solve the vehicle routing problem with stochastic demands (VRPSD). The strategy can realize reoptimization policy under cooperation between a pair of vehicles, and it can be applied in the multivehicle situation. The PCR repeatedly triggers communication and partitioning to update the vehicle assignments given real-time customer demands. We present a bilevel Markov decision process to model the coordination of a pair of vehicles under the PCR strategy. We also propose a heuristic that dynamically alters the visiting sequence and the vehicle assignment given updated information. We compare our approach with a recent cooperation strategy in the literature. The results reveal that our PCR strategy performs better, with a cost saving of around 20–30%. Moreover, embedding communication can save an average of 1.22%, and applying our partitioning method rather than an alternative can save an average of 3.96%.