The market competition being increasingly fierce, the two major traditional sources of enterprise's profits are shrinking. Therefore, the impact of logistic costs becomes more important to enterprise's profits, especially the distribution cost, which occupies a larger proportion in the logistics costs. The vehicle routing problem has larger impact on the distribution cost comparing with other problems; therefore, people pay more and more attentions to the vehicle routing problem(VRP). Particularly from the late 1990s, more scholars and enterprises realize that the importance of the vehicle routing problem in domestic. Based on the analysis of vehicle routing problem, this paper studies the multi-objective model of the vehicle routing problem with time windows. There are four main parts in this paper:Firstly, the paper analyzes a lot of literatures about the vehicle routing problem, and finds out the problems in both the model of VRP and the algorithm for solving the VRP. Furthermore, the paper introduces the basic model of VRP and systemically concludes its types and common algorithms for the VRP.Secondly, there are two main problems in the current research of VRP model. One is the singleness of the objective. The other is that there are some differences between the constraints and the reality. The paper establishes a multi-objective mathematical model of VRPTW by introducing a new objective function to express the satisfaction degree of customers and adding the multi-type vehicles constraint to solve the two problems.Thirdly, the paper uses genetic algorithm to solve the multi-objective mathematical model. To improve the performance of the genetic algorithm, the paper selects better selection operator, crossover operator and mutation operator, and thus avoids the "premature" of the genetic algorithm at a certain extent.Fourthly, the paper uses MyEclipse software to design a programming of the designed genetic algorithm and selects data from two literatures to calculate them. By comparing with the results of the two literatures, the paper ensures the designed effectiveness of algorithm and puts forward some suggestion. |