Using Genetic Algorithms for Multi-depot Vehicle Routing

Efficient routing and scheduling of vehicles has significant economic implications for both the public and private sectors. Although other variants of the classical vehicle routing problem (VRP) have received much attention from the genetic algorithms (GAs) community, we find it surprising to identify only one GA in the literature for the fixed destination multidepot vehicle routing problem (MDVRP). This paper aims to bridge this gap by proposing an application of genetic algorithms approach for MDVRP. The proposed GA employs an indirect encoding and an adaptive inter-depot mutation exchange strategy for the MDVRP with capacity and route-length restrictions. The algorithm is tested on a set of 23 classic MDVRP benchmark problems from 50 to 360 customers. Computational results show that the approach is competitive with the existing GA upon which it improves the solution quality for a number of instances. A comparison of the GA’s approach with other non-GA approaches show that although GAs are competitive for the MDVRP, there is still room for further research on GAs for MDVRP, compared to Tabu search.