摘要
为了解决企业废旧产品回收再循环的问题,提出了一个考虑设施扩展的多周期、多产品、多级的闭环物流网络模型;以经济成本最小和网络响应最快为优化目标对网络中的回收中心、分销中心、再制造中心、再循环中心等设施进行选址和流量分配。为了求解这一多目标优化问题,用二维数组来编码染色体,采用基于进化算法和贪婪算法的两阶段算法进行求解。最后通过算例数值计算,求出了模型解集的Pareto前沿面,并与采用的约束法的CPLEX软件计算结果进行对比,结果表明在小规模问题上两阶段进化算法与CPLEX性能接近,随着问题规模增大,两阶段算法与CPLEX相比,结果误差范围在8.24%以内,但是效率高近10倍。
With facility expansion, a multi-period, multi-product, multi-stage closed-loop logistics net- work model for used product collection and recycling is established. To minimize the total cost and maxi- mize the responsiveness of the network, the location and allocation of the facilities in the network is imple- mented. A two-dimension array chromosome is adopted to represent the decision variables, and a two-stage procedure by using evolutionary and greedy algorithm is proposed to solve the multi-objective optimization problem. At last, a serial of numerical experiments is carried out to find the Pareto Solution Set and verify the proposed algorithm. Result shows that the performance of these two methods are nearly the same for small-scale problems, and the two-stage evolutionary algorithm is 10 times faster than CPLEX with a gap less than 8.24% when the scale of problems increases.
出处
《工业工程》
北大核心
2013年第5期53-61,共9页
Industrial Engineering Journal
关键词
闭环物流
进化算法
设施选址
多目标优化
closed-loop logistics
evolutionary algorithm
facility location
multi-objective optimization
