基于遗传模拟退火算法的多层设施选址方法  

A Method of Multi-echelon Facility Location Based on Genetic Simulated Annealing Algorithm

作　　者：李波[1] 曾成培[1] 

机构地区：[1]天津大学管理学院,天津300072

出　　处：《计算机仿真》2008年第5期252-255,301,共5页Computer Simulation

基　　金：国家自然科学基金资助(70572045);新世纪优秀人才支持计划资助(NCET-06-0236)

摘　　要：逆向物流网络是逆向物流系统高效运作的基础和前提,而设施的选址定位是逆向物流网络设计的核心问题。为此,提出一个多层设施选址模型,旨在构建由回收点、回收中心和生产点相结合的最佳逆向物流回收网络。根据模型特点,提出基于遗传模拟退火算法的求解方法,个体采用二进制十进制混合编码;提出基于Metropolis准则的特定遗传进化操作;设计顾客对回收点、回收点对回收中心的两个子分配算法保证所有约束的满足性。最后通过仿真实验,得到满意的设施选址方案。可见,选址模型和算法是一种有效的设施选址方法,具有一定的应用前景。Reverse logistics network is the foundation for the efficient operation of reverse logistics system, and facility location is the core of the design of reverse logistics network. So, this paper develops a multi-echelon facility location model to design the best return network linking return points, return centers and manufacturing facilities. According to the character of the model, it proposes a solution based on genetic simulated annealing algorithm, in which, each chromosome consists of binary values and decimal values, and the special evolution operation is based on Metropolis rule. In order to satisfy all constraints, two sub- distribution algorithms for dealing with customer-return points and return points-return centers are designed. Finally, satisfactory facility location result can be obtained by simulation experiment. The result shows that the model and algorithm are effective and will have a preferable application future.

关 键 词：多层设施选址 遗传模拟退火算法 回收网络 

分 类 号：O221[理学—运筹学与控制论] TP391.6[理学—数学]