多目标产品配置优化研究  

Research on Optimization of Multi-Objective Product Configuration

作　　者：詹钧凯 石宇强[1] 陈柏志[1] 蔡跃坤 ZHAN Jun-kai;SHI Yu-qiang;CHEN Bai-zhi;CAI Yue-kun(School of Manufacture,Southwest University of Science and Technology,Sichuan Mianyang621010,China)

机构地区：[1]西南科技大学制造科学与工程学院,四川绵阳621010

出　　处：《机械设计与制造》2020年第8期40-44,共5页Machinery Design & Manufacture

基　　金：四川省教育厅资助科研项目(18zd1118)。

摘　　要：针对多目标产品配置优化问题,考虑实例关系和个性化等约束,构建了以性能、成本和交货期为目标的产品配置模型。设计了一种改进的非支配排序遗传算法(Non-Dominated Sorting Genetic AlgorithmⅡ,NSGA-Ⅱ)进行配置模型求解,并根据顾客偏好推荐配置方案。该算法采用动态罚函数处理约束问题,采用自适应交叉和变异概率提高算法收敛速度,对变异操作结果进行模拟退火操作,避免了算法陷入局部最优解,并针对多目标问题改进了Metropolis准则。通过算法验证与实例应用,证明本模型有效可行,改进NSGA-Ⅱ算法在配置问题求解上优于NSGA-Ⅱ算法。Aiming at the multi-objective product configuration optimization problem,considering the constraints of instance relations and personalization, a product configuration model with performance, cost and delivery time as target was constructed. An improved Non-Dominated Sorting Genetic Algorithm Ⅱ was designed to solve the configuration model,and recommend the configuration scheme according to the customer preference. The dynamic penalty function was used to deal with the constraint problem in this algorithm,using adaptive crossover and mutation probability to improve the convergence speed of the algorithm. Simulated annealing was performed on the result of the mutation operation to avoid the algorithm getting into the local optimal solution,and improved the Metropolis criterion for multi-objective problem. The model was proved to be effective and feasible through the verification of the algorithm and the application of the example. The improved NSGA-Ⅱ algorithm outperforms the NSGA-Ⅱ algorithm in the configuration problem solving.

关 键 词：多目标优化 产品配置模型 改进的NSGA-Ⅱ 动态罚函数 自适应 模拟退火算法 

分 类 号：TH16[机械工程—机械制造及自动化] TP18[自动化与计算机技术—控制理论与控制工程]