Real-time vehicle routing: Solution concepts, algorithms and parallel computing strategies

In real-time fleet management, vehicle routes are built in an on-going fashion as vehicle locations, travel times and customer requests are revealed over the planning horizon. To deal with such problems, a new generation of fast on-line algorithms capable of taking into account uncertainty is required. Although several articles on this topic have been published, the literature on real-time vehicle routing is still disorganized. In this paper the research in this field is reviewed and some issues that have not received attention so far are highlighted. A particular emphasis is put on parallel computing strategies.