摘要
为辅助工程师进行产品服务系统方案设计,提出一种基于离散布谷鸟搜索算法与Pareto结合的配置优化方法。以成本、质量和时间为目标函数,采用动态惩罚函数方法将约束优化问题转化为多目标优化问题。基于Sobol序列初始化,采用十进制编码和非支配更新方法改进多目标离散布谷鸟搜索算法,以提高初始解的多样性与优化性能。将所提方法应用于某数控机床产品服务系统方案配置优化,通过结果分析及性能对比,验证了该方法解决高维度空间内产品服务系统方案配置优化的有效性与可行性。
To assist engineers to design the product service system,a kind of configuration optimization method combined with Discrete Cuckoo Search Algorithm（DCS）based on Pareto was proposed.The cost,quality and response time were set as the objective functions,and the dynamic penalty function was adapted to convert the constrained optimization problem into multi-objective optimization problem.Based on Sobol sequence initialization,a decimal coding combined with the non-dominated updating method was used to improve Multi-Objective Cuckoo Search（MOCS）for promoting the initial solutions diversity and optimization performance.Through the comparison of results and performance analysis,the feasibility and validity of the proposed method for solving optimization configuration of product service system in the high dimensional space were verified.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2017年第8期1774-1786,共13页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(51205242
51405281)
上海市科技创新行动计划重点资助项目(16111106402)~~
关键词
产品服务系统
配置优化
Sobol序列
多目标优化
离散布谷鸟搜索算法
product service system
configuration optimization
Sobol sequence
multi-objective optimization
discrete cuckoo search algorithm
