Research On Heterogeneous Vehicle Routing Problem With Soft Time Windows Based On Milk-Run

With the rapid development of automobile industry in China, automobile price has being dropped drastically and profit margin has being shrunken. How to reduce production cost of automobile and maintain market competitiveness become a major problem faced by automobile manufacturers. As a source part of automobile logistics, inbound logistics not only directly affect the overall cost of automobile logistics system, but also concern the cost and quality of vehicle production subsequently. Then, studying how to reduce the cost of inbound logistics is very necessary and important.The application status of milk-run inbound logistics and research situation of vehicle routing problem in domestic and foreign are compared. A mathematical model which is more in line with actual operation about heterogeneous vehicle routing problem with soft time windows based on milk-run is established. The objective function of the model is to minimize the total cost which includes vehicle fixed starting cost, traveling cost and time penalty cost; the constraints are focusing on considering soft time windows and the limited loading capacity of vehicles; and the principle of vehicle scheduling is to minimize vehicle mileage, and maximize pickup punctuality and loading rate.Algorithms about vehicle routing problem are discussed, and genetic algorithm is selected to solve the building model. According to integer coding of suppliers and estimating vehicle numbers, an initial population is generated randomly, the best individual preservation and local competition selection strategy are used in order to accelerate the convergence rate of the population; and reversion, exchange and insertion variation are used for the best one selected to maintain the diversity of group and prevent precocity in the algorithm. Finally, taking industrial park of SQ automobile manufacturer as an example, there are twenty supplies in the industrial park and three vehicle types in milk-run, genetic algorithm with Matlab program is designed, five best pick-up routes of milk-run during industrial park are solved. Comparing with milk-run and supplier delivery directly, the transportation distance is shortened by 50.3%, the number of transportation vehicles is decreased by 75%, the average loading rate is increased by 62.84%, and the total transportation cost is saved by 61.8%. It shows that the established model and designed algorithm are correct and effective for solving heterogeneous vehicle routing problem with soft time windows based on milk-run.