Research On The Vehicle Scheduling Optimization Of The Logistics Distribution

Distribution is an operation linking with consumer directly, and takes account for considerable proportion in variable costs in logistics. The planning of vehicle routing in distribution will take great effect on the efficiency, cost and benefit, especially in distributing for multi consumers. So, vehicle routing problem had become focus of many scholars to study. In the developed commercial society, requirement of consumer for delivery time is higher and higher so that delivery day formerly had turn to delivery hour now. With popularization of Internet and development of electronic commerce, traditional VRP algorithm have been unable to meet the requirement that Quick Response to customer demand had brought forth, then the conception of Time Window has come into being. The vehicle-scheduling problem with time window is also a NP-hard problem being more complicated than VRP. Most scholars have been bending themselves to study heuristic Algorithm that can resolve large-scale problem, but cannot ensure the quality of the resolution.A mathematic model with time penalty function for the vehicle-scheduling problem built in this paper is divided into two sub-problems of clustering and routing. Firstly, Genetic Algorithm is used to cluster the customers in the clustering sub-problem. Then Saving Algorithm is to find the best route of vehicle for each customer cluster, and the routing results are fed back to clustering problem as the fitness of evolution. Finally, a satisfied result could be found through continued iterations of the process. It is indicated that the two-phase hybrid genetic algorithm performs well in both resolving efficiency and result quality, especially to large-scale vehicle scheduling problem with potential of applying in practice to some extend.