Based On Improved Particle Swarm Optimization For Study Of Selection Of Distribution Center

Nowadays, with the rapid development of computer science, the expansion of human living space, as well as the widen about the region of knowing and reforming the world,people have a new and higher requirements for science and technology,especially optimization techniques and intelligent computing. Swarm intelligence, as an emerging smart computing technology is more and more concerned by researchers. Collaborative , distribution, robustness and fast characters of swarm intelligence provided the basis of finding the solution to the complex distributed problem, in the absence of centralized control and is not provided under the premise of the overall model. Particle Swarm Optimization is a new intelligent optimization algorithm, it origined from birds preying simulation system. The algorithm is easier to implement than the genetic algorithm (GA), there is no crossover and mutation operations, the parameters which are needed to adjust is small, and fast convergence, so it has been widespread concerned since kennedy and Eberhart proposed it in 1995. especially in recent years, particle swarm algorithm is more and more favored by the specialists and scholars of swarm intelligence research, and the algorithm has been successful in some areas of applications. Now it is widely used in function optimization, dynamic environment optimization, neural networks training and many other areas.Logistics distribution center play an important role in modern commodity circulation, it can greatly reduce the labor intensity of operations, reduce the consumption of goods, improve the turnover rate of inventory, speed up the circulation of commodities, lower circulation costs, increase satisfaction of social needs, and give more choices to consumers by the unified management of the work, such as transport, storage, loading and unloading,handling, distribution processing, delivery, and the processing of purchase order and information etc. In the logistics network, the distribution center connects the supply point and demanding point, it is a bridge between them, and plays an important role in the logistics system , do a good job in the selection of sites for distribution centers will have a significant impact on the logistics systems and the improvement of logistics economic efficiency. So the study about the current distribution center becomes hot in the logistics development. The location problem is not only the first, but also the most complex problem in the distribution center planning.In this paper, our work includes the following aspects:1. Combine particle swarm optimization with immune algorithm, bring immune operator such as immune memory and clonal selection etc into particle swarm optimization, and develop a developed algorithm. The simulation results show that the algorithm is able to overcome the shortcoming of lower precision, easy-divergence in the basic particle swarm optimization algorithm, speed up the search process, and be avoid of falling into local optimum.2. The location problem in the logistics distribution center is studied, propose thedistribution center location model based on the time and cost, enable the location problem in the logistics distribution center to have more meaning and guidance, and provide reference to the future construction of distribution centers.3. The improved particle swarm optimization is applied to discrete points location problem, and solve the proposed location model.