A Glowworm Swarm Optimization Algorithm for the Multi-Objective Energy Reduction Multi-Depot Vehicle Routing Problem

Abstract

In this chapter, we present, initially, a literature review of the most recent publications for the glowworm swarm optimization (GSO) algorithm. Then, we propose a modified version of the algorithm, the parallel multi-start multiobjective glowworm swarm optimization (PMS-GSO) algorithm, which is suitable for multi-objective routing problems. The proposed PMS-GSO was used for the solution of a very complex vehicle routing problem, the multi-objective route-based energy reduction multi-depot vehicle routing problem (MERMDVRP). We solved four different versions of the problem, from the simplest one, where we had a problem without real-life constraints, to the most difficult one, where a number of real-life constraints based mainly on the behavior of the driver, the road, and the weather conditions have been added. Finally, we compared the computational results of the proposed algorithm with the computational results of three other algorithms, the parallel multi-start non-dominated sorting genetic algorithm II (PMS-NSGA-II), the parallel multi-start multi-objective firefly (PMS-FIREFLY) algorithm, and the parallel multi-start multi-objective particle swarm optimization (PMS-NSPSO) algorithm in a number of benchmark instances.