考虑不确定性的稳健优化设计研究

摘要

传统的稳健设计仅能处理含有随机不确定性参数的问题,以响应的均值和方差来评价.稳健优化设计同时考虑随机不确定性参数和认知不确定性参数,实现不确定性传播的基础上对问题设计参数的优化.采用证据理论来统一描述随机不确定性和认知不确定性,并介绍了描述不确定性时的不确定性传播和响应原理,提出了响应评价准则.最后,以测试系统为研究对象,采用证据理论结构进行不确定性传播,以提出的准则为判断标准,对优化过程进行了验证.%Traditional robust design can take a-leatory uncertainty into analysis, and the average and variance of system response are for evaluation. Robust optimal design can take account of aleatory and epistemic uncertainty, and the optimization of design parameters are done with the transfer of uncertainty. The evidence theory has been used to describe aleatory and epistemic uncertainty uniformly, and transfer and response with evidence theory has been introduced,also the principle of response evaluation has been set up. Finally,the test system has been chosen for analysis, and the optimization process has been validated after the uncertainty transfer and estimation with evaluation principle.