稳健设计中的稳健可行性分析

针对稳健设计中的约束可行稳健性问题,提出了一种新的稳健优化设计方法。通过分析不确定因素对约束的影响,提出了利用最大波动分析计算在不确定性因素影响下约束的最大波动量,将该变化量作为惩罚项添加到原约束中,构造了两级优化数学模型。顶级优化是在原有常规优化的数学模型基础上添加了稳健可行性的判断指标,次级优化用来判断稳健性指标的值。与其他方法比较,该方法不需要知道不确定因素的概率分布和约束的梯度信息。实例结果证明该方法是可行的。

Feasibility robustness analysis in robust optimization designs

A deterministic method was developed to analyze the robustness of constraints in robust designs.The method first analyzed the variations of the constraints by considering the uncertain factors and then quantified the robustness of the constraints using maximum variation analysis.A two-level optimum mathematical model was constructed by adding the maximum variations to the original constraints as penalty functions.Top-level optimization was used to solve the original mathematical model with lower-level optimization used to judge the robustness of the constraints.This method does not require a presumed probability distribution of the uncertain factors or constraint gradient information.Example results show that the approach is feasible for robust design.