Multi-Agent System for Solving the Vehicle Routing Problem: A Hybrid Metaheuristic Approach

Abstract

This paper introduces a novel Multi-Agent System (MAS) designed to solve the Vehicle Routing Problem (VRP), a well-known optimization challenge in logistics. The proposed MAS integrates seven established metaheuristics—Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), Golden Ball Algorithm (GBA), Hill Climbing (HC), Tabu Search (TS), and Simulated Annealing (SA)—using a Bi-directional Communication strategy. This strategy enables the system to dynamically switch between population-based and single-solution-based metaheuristics, effectively balancing exploration and exploitation of the solution space. Experimental results demonstrate that the MAS outperforms traditional methods in terms of solution quality, achieving statistically significant improvements. However, the computational efficiency of the MAS requires substantial further optimization for practical real-world deployment. The proposed framework advances the field of metaheuristic optimization for complex VRP variants, indicating a promising direction for future research.