A comparative analysis of meta-heuristic solutions to vehicle routing problem with time wi̇ndows

A major part of the field of Operational Research is dealing with Vehicle Routing Problem (VRP). VRP identifies the problem of delivering the desired service or goods to the customers of the transportation vehicles used in a wide variety of sectors by using the developing distribution networks. VRP with Time Window (VRPTW) examines the problem of realizing the service or goods to be delivered to the customers in a more realistic scenario within the time period determined by the customer. Many of the current problems facing OR are complex computationally demanding resources. Therefore, it becomes very difficult to find a definite solution to such problems. In such cases, heuristic or meta-heuristic methods are used to find the most approximate solution. Meta-heuristics are evolutionary algorithms that are often inspired by life forms in nature. In this study, an analysis of Genetic Algorithm (GA) and Set-based Particle Swarm Optimization (S-PSO) meta-heuristic methods are experimented on solving the VRPTW problem. The stochastic nature of S-PSO and greedy initialization scheme for GA are examined to report the feasibility of those two techniques in comparison to Solomon Benchmark datasets will be made. Keywords: VRP, meta-heuristics, particle swarm optimization, genetic algorithm