Genetic Algorithm to the Split Delivery Vehicle Routing Problem

Abstract

Determining the distribution routes is important in the field of logistics. In this paper, the split delivery vehicle routing problem (SDVRP), which is a variant of the traditional vehicle routing problem (VRP), is studied. In the research of SDVRP, the demands of each customer are allowed to split and be served at least one vehicle. A mathematical model based on SDVRP, aiming to minimize the total distribution cost, is established. Because the SDVRP is still an NP-hard problem, it is impossible to get optimal solutions in a short time. Therefore, the genetic algorithm is adopted to solve this problem. By applying the genetic algorithm to a practical case and comparing the results before and after optimization, a conclusion indicates that the distance between distribution center and the customer affects the effect of splitting deliveries. As the scope of the split customer increasing, the total cost shows a trend of decreasing.