Advanced Multi-Stage Local Search Applications to Vehicle Routing Problem with Time Windows: A Review

In many applications of the vehicle routing problem with time windows(VRPTW), the primary objective is defined as the minimization of the number of routes or vehicles. Minimization of the total travel cost is introduced as a secondary objective. Recently, it has been recognized that minimizing the number of routes and the travel cost concurrently often drives a search toward solutions with low travel cost, but not necessarily with small number of routes. This paper presents a survey of the latest research motivated by this recognition. The presentation is focused on multi-stage applications of advanced local search techniques on the VRPTW. Multi-stage algorithms optimize the number of vehicles and travel time independently in order to ensure that the search is directed towards the achievement of the primary objective. Basic features of these algorithms, as well as hybridization strategies are described. For most algorithms, experimental results on Solomon’s benchmark test problems are provided and analyzed.