多目标选择性拆卸序列优化问题的分散搜索算法  

Scatter search algorithm for multi-objective selective disassembly sequence optimization problems

作　　者：郭希旺[1] 刘士新[1] 王大志[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110819

出　　处：《系统工程学报》2016年第3期307-316,共10页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(71171038);中央高校基本科研业务费资助项目(N110404024)

摘　　要：针对多资源约束下顺序依赖的选择性拆卸序列优化问题,建立以最大拆卸收益和最小拆卸时间为优化目标的多目标数学模型,提出了一种多目标分散搜索优化算法进行求解.该算法针对本文问题的特点设计了一种保持足够多样性的初始解生成方法,满足拆卸优先关系的交叉组合算子以及改进的参考集更新策略.为了进一步提高解的质量设计了一种局域搜索策略,并利用外部存档方法存放pareto解集.应用多组实例进行计算实验,并与其他求解该问题的算法进行比较,实验结果表明本文算法优于对比算法,证明本文模型和算法求解本类问题有效.Selective disassembly sequence optimization problem with multi-resource constraints and sequencedependence is considered by establishing a multi-objective model with the objective of maximizing profit and minimizing disassembly time.A multi-objective scatter search（SS） method is put forward to solve this problem.Based on the characteristics of the problem,an initial solution diversification generation method is put forward which satisfies the crossover combination operator,which maintains the disassembly precedence relationship,and improved reference set updating strategy.In order to further improve the quality of the solution,a local search strategy is designed and external archive method is employed to store Pareto solution set.The results of SS and other algorithms in solving several computational experiments are compared,and the experimental results demonstrate the validity of the model and SS algorithm.

关 键 词：选择性拆卸序列优化 多资源约束 多目标 分散搜索算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]