一种考虑参数相关性的可靠性优化设计方法

Reliability-based Design Optimization Considering Correlated Random Variables

在求解具有非线性相关性结构的可靠性设计优化问题时,现有的相关性处理方法在某些情况下可能会得到精度较低的优化结果。基于Copula函数提出了一种可靠性优化设计方法,用于求解存在非线性相关性结构的可靠性优化设计问题。该方法根据已知的样本对各备选Copula函数进行参数估计,并根据AIC准则选出最优Copula函数。根据最优Copula函数,构建联合概率分布函数,最后求解结构可靠性优化设计问题。通过两个数值算例验证了该方法的有效性,并讨论了不同Copula函数对优化结果的影响。

When solving structural reliability-based design optimization problems with nonlinear cor-relation,current method for correlated variables might lead to inaccurate optimization results in some sit-uations.Based on Copula function,this paper proposed a method to serve as an effective tool for structur-al reliability-based design optimization problems where nonlinear correlation existed.The proposed method estimated the parameters of alternative Copula functions according to known samples and select-ed the optimal Copula function by AIC criterion.Therefore,joint distribution function of variables was established to solve structural reliability-based design optimization problems.Finally,two numerical problems were used to demonstrate the validity of proposed method.Influences on optimization results caused by different Copula were also discussed.