The Genetic Algorithm Applied Research In The Freight Logistics

Besides to reduce cost and to improve labor productivity, modern physical distribution has been the third source of profit for corporations. The science of physical distribution has been one of the most influential new sciences. The physical distribution and electronic commerce is developing very fast. The traffic of goods sending is increasing very fast, too. These developments promote the advance of physical distribution management. Accordingly, management with information has been more and more important.Vehicle routing problem is a typical NP-hard problem. Heuristic algorithms were put forward for solving NP-hard problems. Modern optimizing algorithms are a genus of heuristic algorithms. Genetic algorithm is a kind of modern optimizing algorithm. It is effective to solve VRP. Genetic algorithm is a random search algorithm, using the reference of natural selection and genetics theory in living nature. Genetic algorithm has been applied to plenty of search process in scientific research and projects. It can effectively resolve the optimizing problem of combinations and complex functions, which is hardly possible to conventionality optimizing methods.We researched physical distribution problem of highway transport proceed from Zhengzhou Information Center of Highway. Based on the VRP model, we designed a concrete, practical mathematical model. We realized every step of the algorithm for resolving the model. We coded freight array with natural number. We got loading scheme and transport channel through decoding it with vehicle array. We produced original population with random way. We worked out the degree of adaptability by calculating the schemes, which decoded from these units. We designed a function of return particularly. We adopted the way of roulette wheel to realize the select step. We adopted the way of single discrete cross to realize the cross step. We adopted the way of single even mutation to realize the mutate step. These steps produced child units. The algorithm finished by time standard and offered a optimum solution.We examine the algorithm with classical example and actual example. We found the search extent of the algorithm is too narrow. The algorithm converges local easily. We improve the algorithm at the cross step. We adopted the way of area recombine cross. The compare data of experiment proved the effect of the improvement. We...