The VRP is a well known combinatorial optimization problem. Customers require simultaneous pick-up of goods from their location in addition to delivery of goods to their location in some cases. The VRP-SPD is an extension of the CVRP and considers simultaneous distribution and collection of goods to/from customers. A fleet of vehicles originated in a depot serves customers with pick-up and deliveries from/to their locations.
VRP-SPD has been receiving growing attention due to the increasing importance of reverse logistics activities. Also VRP-SPD has a combinatorial nature. Due to GA’s efficiency in solving complex optimization problems, a GA based approach for solving VRP-SPD was proposed in this study.
Feasibility issue is more complex and hard to be ensured in VRP-SPD than in VRP-B due to the simultaneous delivery and pick-up activities as mentioned in Section 2. This study contributes to the VRP field by providing an efficient and effective GA based approach that produces highly feasible routes for VRP-SPD. Details of the proposed approach were given in the previous sections, after introducing VRP-SPD and its mathematical formulation. Then, an illustrative example and parameter settings were presented followed by performance evaluation of the proposed approach with computational experiments. Twenty four CVRP test problems were transformed into VRP-SPD problems by adding pick-up demand data randomly to the selected problems. According to the results of computational experiments, it can be concluded that the proposed GA based approach both performs well and is efficient. The results of computational experiments prove that, the solutions of the proposed approach have both weak feasibility and strong feasibility. While decoding procedure guaranteed weak feasibility of solutions, fitness value enabled the proposed approach to produce strongly feasible routes.
Direct genetic representation was used for encoding in this study. Further studies may use indirect representation methods such as random key based representation to improve solution quality and efficiency of the procedure. Additionally, the proposed approach may be applied to a real world routing problems with simultaneous pick-up and deliveries.