非完整信息条件下的结构可靠性鲁棒设计

在非完整信息条件下,建立区间变量、模糊变量、随机变量同时存在的可靠性设计模型,并给出相应的序列优化设计方法;同时考虑设计变量的实现值与设计值之间的差别以及约束条件的鲁棒性,提出基于info-gap决策理论的鲁棒优化模型,研究可靠性鲁棒设计的序列求解策略。算例1表明:模糊可靠性设计比随机可靠性设计结果保守,当模糊设计指标α=0时(对应区间可靠度)结果最为保守。算例2表明:不同的鲁棒性指标αt对应不同的鲁棒优化解,当αt大于0.20时,不存在相应的鲁棒优化解。利用本文提出的鲁棒序列优化求解方法对目标函数计算1291次即可,而常规优化算法需8107次,表明本文算法大大提高了计算效率。

Reliability robust design of structures under incomplete information

This paper first discusses the reliability-based design optimization(RDO) of structures under the mixture of random,interval and fuzzy variables.The sequential single-loop optimization method is applied for RDO with different combinations of uncertaities.Info-gap models are used to represent errors between "true" values and design results.A sequential algorithm for the robust reliability design is presented.Example 1 shows that the fuzzy reliability design can offer conservative result.The results of example 2 show there are different robust optimal solutions when the robustness index takes different values.When α t is greater than 0.2,the robust optimal solution does not exist.After 1291 times computation of the objective function,the optimal result can be obtained by using the sequential algorithm for the robust reliability design,while 8107 times computation are needed by the conventional optimum algorithm.