基于两类因子波动下的稳健优化设计  

Robust Optimization Design Based on Two Types of Factors Fluctuation

作　　者：韩云霞[1] 马义中[1] 欧阳林寒 刘丽君 顾晓光[1] 张延静[1] 

机构地区：[1]南京理工大学经济管理学院,江苏南京210094 [2]南京航空航天大学经济与管理学院,江苏南京210016

出　　处：《工业工程与管理》2017年第6期25-31,39,共8页Industrial Engineering and Management

基　　金：国家自然科学基金资助项目(71471088;71371099);中央高校基本科研业务专项资金资助项目(3091511102);江苏省自然科学基金资助项目(BK20170810)

摘　　要：针对稳健优化设计中噪声因子和可控因子波动对质量特性影响的问题,结合Kriging模型和稳健优化思想,提出了一种同时考虑噪声因子和可控因子波动的稳健优化方法。该方法首先假定噪声因子和可控因子服从正态分布,提出了基于两类因子波动下的稳健设计方法;其次,在融合两类波动信息的空间填充设计基础上,构建Kriging均值模型和标准差模型;然后,基于稳健优化思想,建立优化目标函数;最后,结合经济订购批量(Economic Order Quantity,EOQ)模型和蒙特卡洛仿真,验证所提方法的有效性和稳健性。结果表明,所提方法能有效解决噪声因子和可控因子波动情况下的稳健优化问题。Noise factors and controllable factors exert some influence on quality characteristic. A new robust optimization method was proposed based on Kriging model and robust optimization ideology. Firstly, based on the hypothesis that the noise factor and the controllable factor follow normal distribution, a robust optimization method was presented regarding the variation of two types of factors. Secondly, Kriging mean and variance models were constructed by integrating two kinds of fluctuation on basis of space-filling designs. Then optimization objective function was developed following robust optimization idea. Finally, the economic order quantity model and Monte Carlo simulation were combined to validate the effectiveness and robustness of the proposed method. Simulation results indicate that the proposed Kriging method can solve the problem of robust optimization under the fluctuation of noise factor and controllable factor.

关 键 词：噪声因子 可控因子 波动 KRIGING模型 稳健优化设计 

分 类 号：F273.2[经济管理—企业管理]