Harmony Search for Multi-depot Vehicle Routing Problem

This study considers the multi-depot vehicle routing problem in supply chain network design. Vehicle routing is one of the important system in supply chain. The objective is to find the minimum distance travelled by the vehicles, from the depot to the customers. We proposed an improved harmony search algorithm for solving this problem. Firstly, the Clarke & Wright saving algorithm is used for the initialisation of a solution vector in harmony search. Three simple heuristics approaches; swapping, insertion and relocation are proposed as the local optimisation techniques during the implementation of the harmony search. The roulette wheel approach is implemented for the selection procedure. Computational experiments are conducted using the benchmark dataset of Cordeau's problem instances. Computational results show that the proposed harmony search algorithm is comparable to other metaheuristic approaches from the literature.