Application Of Improved Genetic Algorithm To Logistics Vehicle Scheduling Problems

With the advent of the information era, E-commerce (E-business) based on the information technology has sprung up throughout the world, which has a wide and deep influence over the traditional modes of the business operation and the commodity circulation. As"the Third Profit Resource", logistics has been an important component of the commodity circulation. Having an increasingly visible impact on economic activities, logistics is growing to be a critical domain of competitiveness.Logistics, as a comprehensive industry in the end of the 20th. century, is becoming more and more important in the development of national economy and in the rising of managing level of enterprise management. To make logistics perform full function, and to make it perfect and optimized constantly, it is necessary to make logistics rationalized. Logistics systematization is the key to make logistics rationalized, so logistics must take the road to systematization. Therefore, enterprises must strengthen the construction and application of logistics system in order to improve their synthesized competitiveness.Logistic distribution is an operation linking with consumer directly, and takes account for considerable proportion in variable costs in logistics. The planning of vehicle scheduling in distribution will be take great effect on the efficiency, cost, and benefit, especially in distributing for multi consumers. The long time and high cost of logistic allocation have become the"bottle-neck"of the expansion of the electronic commerce in our country. A scientific and reasonable method to vehicle scheduling is an important operation in logistic distribution.The research target of the paper is to explore the theories and methods to optimize logistics, reduce the logistic cost, optimize the social stockinstallation, improve the service quantity and sequentially increase economic and social benefit of the business enterprise, and guarantee the healthy development of logistic system with appropriate scale.So, vehicle scheduling problem had become focus of many scholars to study. In the developed commercial society, with popularization of internet and development of electronic commerce, the requirement of consumer for delivery time is higher and higher. So that delivery day formerly had turned to delivery hour now.Now, the problem is not only applied to the field of auto transportation, but also to ship, aviation, communication, electricity, industry management, computer application etc. The algorithm has been applied into many combinatorial optimization problems such as the trainman's shift arrang- ement in aviation , the optimization design of cargo arrangement in ship company, traffic routing arrangement, and the plan and control in the production system.VSP is a typical strong NP-hard problem, high effective exact algorithm is impossible to it. Heuristic Algorithm can resolve large-scale problem, but cannot ensure the quality of the resolution. Recently, genetic algorithm has been tried to resolve variable combinatorial optimization problems, such as job scheduling problem, but has began just now in VSP. Some people assert that genetic algorithm has tended to NP-hard problem.When analyzing logistics system, some difficult systems must take mathematic analytical method as mid-means to solve these problems in the way of model trial. It is generally accepted that the model is a description of actual models, that there may be various modeling in logistics system. Some models and datum are necessary to location of logistics distribution center when we analyze and design logistics system. There are many solutions to the problems of location, but it is difficult to solve the big size system by these orthodox methods in the practical application. Take genetic algorithmto solve and optimize in benefit or function optimization and evaluation analyzing of system or subsystem.Genetic algorithm is a random search and optimization method based on natural selection and genetic mechanism of the living beings. The combination of genetic algorithm and logistics system has good practicability and extensive application prospect.This paper introduces the characteristics and basic principle of genetic algorithm, as well as its prospect, and expounds the primary content of the logistics and the related problem of the logistics system. The paper studies the application method of genetic algorithm for modem logistics system in the distribution problem. Numerical experiments show that the proposed method is more valid and suitable to solve some problems for modern logistics system.Firstly, the dissertation introduced the background and meaning of selecting the topic, elaborated the relationship between electronic commerce and logistics and proposed the conception of building the logistics allocation system of electronic commerce in our country. Afterwards, it theoretically analyzed the two typical problems in the optimization of logistic allocation (the selection of sites and optimization of vehicle routes), established the mathematical model under certain assumption prerequisite and briefly introduced the tool (genetic algorithm) with whole random search capability to solve problems; Finally, it made a series of improvements to the fundamental genetic algorithm in operating aspects of selection, crossover and mutation, applied it to the problem of optimization of logistic allocation and carried out beneficial experiment and analysis.At first this article analyses the factors of distribution logistics center location and sets up its reasonable location mathematics model and discusses arithmetic of mathematics model combining inheritance arithmetic. An improved genetic algorithm (IGA) was presented for the solution to thisproblem. The core of IGA is the PMX (Partially Matched Crossover) operator for a one-point crossover, which overcomes the weakness of genetic algorithm in big size system searching.Based on the analysis of the existing genetic algorithm for the vehicle routing problem (VRP), a distribution model relevant to route arrangement was established, and a mountaineering operator was presented which overcomes the weakness of genetic algorithm and local search algorithm.In order to check the validity of the proposed improved genetic algorithm, the paper worked out the source programs enclosed in the annex and made substantial analysis of practical problems by the program.Through the results of substantial analysis it can be seen that the convergence rate of the whole algorithm can be greatly improved by adopting the improved genetic algorithm. And the chance of the solution falling in local optimization cuts down, the compute velocity and efficiency are improved at de same time. So the improved genetic algorithm is suitable to solve some problems for modern logistics system, thereby the article could provid an effective method to solve the related problems of optimization.