一种逆向物流网络的多期动态选址方法  

Method of multi-period dynamic location in reverse logistic network

作　　者：李波[1] 曾成培[1] 

机构地区：[1]天津大学管理学院物流工程与管理系,天津300072

出　　处：《管理科学学报》2008年第5期76-84,共9页Journal of Management Sciences in China

基　　金：国家自然科学基金资助项目(70572045)

摘　　要：提出一种逆向物流网络的多期动态选址方法.不同于静态的、单期模型,考虑需求变化的不同时期下,计算机相关产品逆向物流网络的多层设施定位方法.基于遗传算法,采用二进制十进制混合编码的染色体来表示回收点、回收点的回收期及回收中心相关的决策变量;设计顾客对回收点、回收点对回收中心的两个子分配算法来保证所有约束的满足性,并进行拟合度函数的评价;提出的特定遗传进化操作使得模型求解方向趋于回收点、回收中心和生产点相结合的最佳逆向物流网络;最后的仿真实例说明了所提方法的有效性.This paper proposes a method of multi-period dynamic location in reverse logistic network. Unlike the static, single-period models, under the different demands in different periods, this paper considers applying a dynamic facility location approach to develop the muhi-echelon reverse logistics network for computer and relevant product returns. Based on genetic algorithm, each chromosome, which consists of binary values and decimal values, denotes the decision variables correlated with return points, return points＇ return period and return centers. In order to satisfy all constraints, we design two sub-distribution algorithms which deal with customer- return points and return points- return centers, and then the fitness function can be evaluated. The special evolution operation proposed in this paper can make the solution of the model tend to be the best reverse logistics network linking return points, return centers and manufacturing facilities. The usefulness of the proposed model and algorithm was validated by an illustrative example.

关 键 词：逆向物流 遗传算法 多期动态选址 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]