ANALYSIS OF PARTICLE SWARM ALGORITHMS FOR THE ELECTRIC VEHICLE ROUTING PROBLEM

Abstract

This paper addresses the application of Particle Swarm Optimization (PSO) and its enhanced version, Generalized Particle Swarm Optimization (GPSO), in solving electric vehicle routing problem (EVRP). The main objective is to compare the performance of PSO and GPSO using different particle representations, specifically: priority vector and adjacency matrix. Experimental results show that the matrix representation significantly outperforms the vector representation, and the GPSO achieves better performance than the basic PSO. Additionally, the paper provides explanations of two algorithms based on the Bellman-Ford algorithm, which serve to divide the complete route into smaller ones and determine optimal positions for charging stations.