• CN：11-2187/TH
  • ISSN：0577-6686
机械工程学报

›› 2010, Vol. 46 ›› Issue (20): 182-190.

• 论文 • 上一篇    

扫码分享

鲁棒协同优化模型及其求解方法

李海燕;马明旭;井元伟   

  1. 东北大学辽宁省复杂装备多学科设计优化技术重点实验室;东北大学流程工业综合自动化教育部重点实验室;东北大学信息科学与工程学院
  • 发布日期:2010-10-20

Model and Solution Methodology Research of Robust Collaborative Optimization

LI Haiyan;MA Mingxu;JING Yuanwei   

  1. Liaoning Province Key Laboratory of Multidisciplinary Optimal Design for Complex Equipment Key Laboratory of Integrated Automation of Process Industry of Ministry of Education, Northeastern University School of Information Science & Engineering, Northeastern University
  • Published:2010-10-20

PDF

151

摘要： 从提高鲁棒协同优化(Robust collaborative optimization, RCO)模型的计算效率和模型求解方法优化性能角度出发，提出简化的基于隐含不确定性传播的鲁棒协同优化(Simplified implicit uncertainty propagation robust collaborative optimization, SIUPRCO)模型和改良的鲁棒协同优化模型求解方法(Improved robust collaborative optimization resolving method, IRCORM)。SIUPRCO是对基于IUP方法所建立的RCO模型进行的改善，它避免对全局灵敏度方程的求解，从而提高了RCO模型的计算效率。在IRCORM模型求解方法中，为解决RCO易陷入局部极值点的问题，利用动态罚函数法得到确定性协同优化的全局极值点，并将该值作为RCO优化的初始点；采用遍历组合的方法，给出Pareto最优解。典型算例的优化结果表明，SIUPRCO模型具有合理性，IRCORM求解方法具有良好的优化性能。

关键词: Pareto最优解, 不确定性, 动态罚函数, 鲁棒协同优化

Abstract: To improve the efficiency of calculation of robust collaborative optimization(RCO) model and the optimization performance of solution method, a new robust collaborative optimization model based on IUP(SIUPRCO) and an improved solution method(IRCORM) are presented. The solution of global sensitivity equation is avoided and the efficiency of calculation is improved in SIUPRCO model. In the initial stage of IRCORM algorithm, to avoid the phenomenon that the optimization results of RCO usually converge to the local extremum, the dynamic penalty function method is adopted to get the global extremum without considering the uncertainty factor and the optimization result is used as the initial point of RCO. The Pareto-optimal solutions are obtained by using traversal combination method. Two typical examples are adopted to test the two presented methods. The results show that the SIUPRCO model is reasonable and the IRCORM method has good optimization performance.

Key words: Dynamic penalty function, Pareto-optimal solutions, Robust collaborative optimization, Uncertainty

中图分类号: 

版权所有 © 2019 《机械工程学报》编辑部 | 京ICP备05055788号-30