ELECTRIC VEHICLE ROUTING PROBLEM USING ADAPTIVE SIMULATED ANNEALING

Abstract

Road transport is a major CO2 emission contributor globally. To tackle the challenge of reducing world carbon emissions, alternative technologies for the automobile industry are widely researched. The automotive industry has started to shift from Internal combustion engine (ICE) vehicles to electric vehicles (EVs), where EVs are the future of the automotive industry in terms of reducing greenhouse gas emissions and air pollution. EV manufacturers are continuously looking for opportunities to optimize the supply chain processes, aiming for supply chain resilience. In this study, we present an Electric Vehicle Routing Problem (EVRP) to achieve the best decision, which is an extension of the traditional Vehicle routing problem (VRP) which in particular finding the shortest route for electric vehicles. The objective function is to find the best travel route that minimizes travel distance. Each route serves a set of customer nodes that starts and ends at a given depot node. We take battery capacity and charging stations as the constraints. In addition, the use of homogenous fleets and single depot are considered in this paper. A hybrid metaheuristic approach is used to find the best solution with the Adaptive Simulated Annealing algorithm. The use of adaptive in simulated annealing generates a higher probability of finding the best operators, which results in better solutions. A comparison of results from various metaheuristic methods is also presented in this paper to get the best method for the EVRP based on a benchmark dataset. This paper ends with recommendations for creating a routing plan that is resilient to disruptions to distribution.