Optimizing the routing and charging station installation location for electric commercial vehicles

Land transport, the most preferred and widely used type of transport especially as it offers door-to-door transportation services, is on its way to becoming a green sector; and today, there is a tendency in this sector towards using vehicles that will be less harmful to the environment. One of these vehicles defined as alternative fuel vehicles is electric commercial vehicles. Although their popularity is increasing day by day, there are some obstacles that prevent them from becoming widespread. Some of these obstacles are the vehicles' limited range in terms of the distance they can go, their long charging time, the low number of charging stations, and the high cost of purchasing these vehicles. This thesis therefore examines two of the main topics of logistics and optimization, i.e., vehicle routing and location selection problems, for electric commercial vehicles both in terms of Electric Vehicle Location-Routing Problem. Differing from other studies, this study aims to determine vehicle routes, charging station locations and types of charging stations for electric commercial vehicles by bringing together constraints such as energy consumption, battery wear and tear, and drivers' working hours. The problem is modeled as a Mixed-Integer Nonlinearly Constrained Optimization Model, and the aim has been determined as minimizing the sum of vehicle purchasing cost, travel cost and vehicle charging cost. In order to solve the problem, the data set developed by Schneider, Stenger, and Goeke (2014) is used. Since the problem is large-sized and contains nonlinear constraints, it cannot be solved using exact solution methods; therefore, the study uses the Genetic Algorithm method, which is a metaheuristic method. Based on an examination of the existing surveys, it is believed that the study will prove successful in contributing to the literature since the problem incorporates many constraints, since no other study has been encountered that solves the problem using genetic algorithm, and since the number of studies on this subject within the Turkish literature is quite low. When we examine the study results, we see that the proposed genetic algorithm offers a solution within a reasonable time, and the energy consumption is less since a realistic energy consumption function is utilized. Furthermore, although the number of vehicles was not limited in the mathematical model, it has been observed that in comparison to the optimization-oriented solution, in most small-sized sample problems, the route can be completed with the same number of vehicles or less. Since the problem allows charging stations to be installed at customers' locations, decisions have been made in this direction when solving the problem, which reduced the completion time of the route. It is observed that mostly fast charging stations are installed in the solution. When similar studies are examined, it is seen that the decision to install fast charging stations in order to meet the time window constraint is compatible with the rest of the literatüre.