论文部分内容阅读
针对基本输入变量概率分布信息不可准确预知的情况,对基本变量基于方差的重要性测度进行分析,建立了随机不确定性分布参数,区间不确定性分布参数以及随机和区间混合不确定性参数条件下的3种重要性测度指标.该指标体系能够反映3种不确定性分布参数条件下,输入变量对结构系统输出性能的影响.由于基于方差的重要性测度计算量大,文中采用稀疏网格方法在每一分布参数实现值处进行计算.针对3种不确定性分布参数情况,分别利用稀疏网格方法和稀疏网格方法结合遗传算法对所提的重要性测度指标进行计算.通过数值算例和工程算例说明所提指标的合理性和所采用方法的高效性.
According to the fact that the probability distribution information of the basic input variables can not be accurately predicted, the variance-based importance measure of the basic variables is analyzed. The random uncertainty distribution parameters, the interval uncertainty distribution parameters and the random and interval mixed uncertainty parameters are established Which can reflect the influence of input variables on the output performance of the structural system under the condition of three kinds of uncertainty distribution parameters.Because the calculation of the importance based on variance is large, Method is used to calculate the value of each distribution parameter.According to the three kinds of uncertainty distribution parameters, the sparse grid method and the sparse grid method are combined with the genetic algorithm respectively to calculate the importance measure index.According to numerical calculation Examples and engineering examples illustrate the rationale for the proposed indicators and the effectiveness of the methods used.