Evolutionary Local Search Algorithm to Solve the Multi-Compartment Vehicle Routing Problem with Time Windows

Vehicle routing problems (VRPs) represent an important research stream in applied combinatorial optimization. As the problems emerging from the practice become more constrained and large, more sophisticated approaches are required to address them. This paper tackles the Multiple Compartment Vehicle Routing Problem with Time Windows. This extension of the classical VRP consists of an unlimited homogeneous fleet of vehicles, each being equipped with multiple compartments. Such a configuration enables to load distinct commodities. A customer is characterized by a nonnegative demand of each commodity and a time window allowing its delivery. The goal is to service all customers under the vehicle capacity constraints and time windows constraints with minimal total cost. A metaheuristic solution approach based on evolutionary local search is presented. The performance of the algorithm is evaluated on benchmarks available in the literature.