Comparison of vehicle routing problem with two-echelon vehicle routing problem and a hybrid model proposition for location selection

It is vital for businesses to carry out their activities profitably and to be sustainable while performing these activities. In this context, multi-stage distribution systems, in which the distribution network is divided into more than one level, can enable enterprises to benefit from both the large amounts of shipments they make in the first stages and the economies of scale arising from the efficient use of vehicle capacities at all levels. On the other hand, multi-stage distribution systems can also reduce the environmental and social impacts resulting from distribution activities carried out in dense urban areas. The first aim of this study is to examine whether multi-stage distribution systems have an advantage over single-stage distribution systems. For this purpose, Capacitated Vehicle Routing Problem (CVRP) and Two-Echelon Capacitated Vehicle Routing Problem (2E-CVRP) were compared according to the key performance indicators of distance, fuel consumption, time and cost. A supermarket chain operating in Turkey in the implementation of the problem was modeled with realistic data distribution network. Sensitivity analysis was conducted with 25 scenarios in order to analyze the effect of changes in some parameters (satellite capacities, demand quantities, number of vehicles in the fleet and vehicle capacities). In addition, single-echelon and two-echelon distribution systems were compared, taking into account the rental and investment costs of the vehicles. Another aim of the study is to propose a method that will enable decision makers to make both location selection and routing decisions efficiently, especially in enterprises that use distribution activities intensively. In this context, contrary to what is usual in the literature, a hybrit method in which the Two-Echelon Vehicle Routing Problem (2E-VRP) and Data Envelopment Analysis (DEA) are used together and many factors (distance, fuel consumption, time, cost) are taken into account in terms of distribution activities was designed and the applicability of the method in question was demonstrated. In addition, this designed method can serve not only for the efficiently selection of potential sites, but also for the purpose of calculating the efficiencies of existing sites.