Study Of Distribution Routing Problem Based On Multiple Objective Genetic Algorithm

With the development of world economy and modern science and technology, as an important service industry of national economy, logistics industry is developing rapidly in worldwide and will increasingly be the crucial and basic industry to promote national economy development. The development of logistics industry has already been one of important marks which are used to evaluate the modernization and the overall national strength of one country.The efficiency of logistics distribution is usually measured by logistics cost. The transportation cost always takes more than half of logistics cost. In the other word, the most effective and directly way to reduce the logistics cost is to reduce the transportation cost. Evidently, the optimization of distribution routing has great sense to logistics distribution. In this paper, many Articles at home and abroad and the features of modern logistics distribution are analized. The study includes three parts surrounding multi-objectives. They are setting up multi-objectives routing models, comparing and choosing solution to the multi-objectives problem and deciding the weight of each objective.Firstly, all kinds of factors effecting distribution routing are considered to help deciding three typical objective functions. Two kinds of distribution routing models with different constraints are set up based to traffic jam and sudden situations. Because the method of searching pareto solution first, then making a decision to choose the optimal solution is very flexible, so the NSGA-II with more efficiency than other multiple objective genetic algorithms is choosed after various solutions to solve multi-objectives problem has been studied. The operators in NSGA-II are made special change to meet the features of distribution routing problem. A group of pareto solutions are get by using NSGA-II, deciders need to choose one solution which satisfies them most. The accuracy of offered information and the preference of deciders can be reflected by the weight of each objective using TOPSIS with the membership degree and entropy. In the end, lianhua supermarket in Hangzhou Jianggan district is taken as an example, the algorithm in this paper is proved using MATLAB programm, and the weights of each objective are determined.