Research On Application Of Improved Particle Swarm Optimization For Joint Distribution

In recent years, with the increasingly fierce market competition, more and more logistics enterprises of small and medium sized participate in the alliance of joint distribution to improve their competitive ability to respond to the external environment. To create an effective alliance of joint distribution needs to solve many problems, of which profit-distribution and distribution centers'location are particularly important. Profit-distribution has been a sensitive topic which many enterprises face, and a fair and equitable mechanism of profit-distribution is a key of an efficient and persistent joint distribution. As well, as a key node in the logistics system, distribution center's location can impact on the cost consumption after cooperation, further to impact on profit-distribution of joint distribution. Whether can these problems be resolved will ultimately impact on the smooth implementation of joint distribution. Because of this, the article recommends distribution centers' location, profit-distribution and improved PSO a useful research. Because the distribution centers' location and profit-distribution are two different issues and they have different emphases, they use different improved methods of PSO.Specifically, the article will systematically study on the following aspects:(1) For the characteristics of joint distribution, set up a location model of joint distribution based on "soft time windows" and "hard time window", expanding the applicable scope of the model further to resovle the question that each customer's requirement of distribution time has to be taken to consideration in the process of cooperation. Because this model is a nonlinear programming problem of multi-constraint and multi-variable, a heuristic algorithm combined PSO with TS was proposed. Simulation results show that putting the tabu search algorithm into the later period of PSO can overcome basic PSO's shortcoming that easy falling into local optimum leads to the results of low accuracy and slow convergence in the late of algorithm.(2) For the most common methods of profit-distribution focus on the direct analyzing the distribution of final profit, but don't take into account members' performance in the whole operation process of the alliance, and ignore the time factor's impact on the whole operation process. In this paper, introduce the time factor and incentives in the problem of profit-distribution, take the program's overall satisfaction and overall net profit as the objective functions and establish a multi-objective optimization model. As multiple objectives usually conflict with each other in the multi-objective optimization problem, there are no optimal solutions for all objectives at the same time. Against this situation, improved PSO and Pareto method which are used to handle multi-objective optimization problems are combined to solve the multi-objective optimization model of profit-distribution, by means of a smulation. Finally, by means of a simulation, verify the effectiveness of the method.