Application Of Genetic Algorithm In Logistics Distribution Of Logistics Enterprise

Logistics industy has developed very fast in recent years. With the logistics market becoming stronger; the competition has become more and more drastic. The high quality of logistics service required by customers and competition in the logistics service make logistics companies pay more and more attention to scientific and efficient logistics distribution management. The development of information and communication technology offers powerful technical support for the efficient and accurate logistics distribution. An important content while studying VRP (vehicle routing problem ) is the modern logistics system, a lot of real transportation logistics and question of providing and delivering can all be expressed as VRP. By choosing suitable transportation route, one can speed up the response to customer's demand, improve the service quality , strengthen customer's satisfaction to logistics system, and reduce the operation costs of the facilitator. According to the customer's demand, VRP can be divided into determinable VRP and undeterminable VRP. Determinable VRP studies a kind of problems in which customer's demand information, quantity and question of geographical position are already known; the latter includes random VRP and fuzzy VRP, in these VRPs, the customer's demand or quantity are indefinite. Almost all the VRPs involve optimazing and programming. The functions in question are often non-smooth. Therefore classical algorithms do not work well for those cases. The heuristic algorithm is an improved search algorithm in the state space. In the algorithm, every position searched is evaluated and a "best position"is selected. From this "best position"the searching continues until the goal is reached. Doing in this way, a large number of useless route searching is avoided. In heuristic searching, the appraisal to the position is very important. There can be different results after adopting different appraisal. The heuristic algorithm can obtain a lot of satisfied feasible solutions to VRP within shorter time. In this article we will use heuristic algorithm to solve VRP. Genetic algorithm is an artifical intelligence technology of self-organization and adaptation, which imitates natural oraganisms'evolutionary process and mechanism to solve problems. The algorithm has been widely applied to computer science, artificial intelligence, information technology and engineering project. The following are outline of works done in this subject: Fistly, the paper introduced the background and meaning of selecting the topic, elaborated the relationship between electronic commerce and logistics and theoretically analyzed the Vehicle Routing Problem, established the mathematical model under genetic algorithm. Secondly, the basic concepts and theory of genetic algorithms are introduced and the then design of genetic algorithm has been discussed, some classical strategies used by genetic algorithm to solve nonlinear programming models will be shown in this chapter as well. A series of improvements are made on the fundamental genetic algorithm in operating aspects of selection, crossover and mutation. An improved method----partheno-genetic algorithm (PGA) is introduced and is used to resolve Vehicle Routing Problem. Genetic algorithms (GA) using ordinal strings must use special crossover operators such as PMX, OX and CX, instead of general crossover operators, such crossover operators are difficult to realize. Considering the above deficiency of GA with using ordinal strings, we proposed a partheno-genetic algorithm (PGA) that uses ordinal strings and repeals crossover operators. We introduced some particular genetic operators such as gene exchange operator which has the same function as crossover operators.Therefore genetic operation of PGA is simple and its initial population needs not to be varied and there is no immature convergence in PGA. Finally, we carried on a real example analysis to the practical problem and have finished the feasible test of the method and function test. The difference between canonical genetic algorithm and rtheno-genetic algorithm in the experiment has verified the superiority of rtheno-genetic algorithm to the canonical genetic algorithms in some class of pratical problems. According to the requirements of Material and Equipment Company of Daqing, we have designed a set of logistics system with which providing and delivering can be carried out. The system is programmed by VC ++ language which has realized this intellectual operation and control in providing and delivering work of company,...