Experimenting crossover operators to solve the vehicle routing problem with time windows by Genetic Algorithms

In the Vehicle Routing Problem with Time Windows (VRPTW), a set of vehicles with limited capacity are to be routed from a central depot to a set of geographically distributed customers with known demands and predefined time windows. To solve the problem, an optimum assignment of vehicles to each customer is needed to achieve the minimal total distance travelled without violating the vehicle capacity and time window constraints. VRPTW has been the subject of intensive research, mainly focused on the design of effective metaheuristics to cope with the high complexity of the problem. Among these, Genetic Algorithms (GAs) have received a special attention, especially for their capabilities in coding problem solutions and achieving good performance on real‐world applications. The goal of this paper is to revise the state of the art and the performance of GAs applied to VRPTW; in particular, we experimented with different crossover techniques and provide a new crossover operator, called DAX, to find high quality solutions. Computational results on the performance of the operators are given on a real‐world application.