论文部分内容阅读
点估计法是随机系统响应量统计矩计算的方法之一,由于简单、高效而颇受关注,其中单变量函数的统计矩估计则是点估计法的基础。虽然各研究者对各自提出的点估计方法均进行了算例验证,但这些算例验证的普适性值得商榷。该文通过详细、系统的研究,对已有的单变量函数统计矩的点估计方法进行全面的影响因素分析和计算性能评价。大量的算例分析结果表明:1)函数的非线性程度、随机变量的类型及其变异系数是点估计算法精度的主要影响因素,变量均值影响较小,且本质上是通过改变函数的非线性程度间接影响精度;2)Zhou&Nowak方法(5个计算点)精度最优;3)当函数非线性程度较强、变量变异系数较大时,各方法精度均不够理想,此时应慎用点估计法。
The point estimation method is one of the methods for calculating the statistical moment of response of random systems. Due to simplicity and high efficiency, the point estimation method is the basis of the point estimation method. Although each researcher validates the proposed point estimation methods, the universality of these examples verification is debatable. Through detailed and systematic research, this paper makes a comprehensive analysis of the influential factors and evaluates the computational performance of the existing method of point estimation of moment of the univariate function. A large number of example analysis results show that: 1) The degree of nonlinearity of the function, the type of the random variable and its coefficient of variation are the main influencing factors of the accuracy of the point estimation algorithm, and the mean of the variable has less influence. By essentially changing the nonlinearity of the function Degree of indirect impact on accuracy; 2) Zhou & Nowak method (5 points) the best accuracy; 3) When the degree of non-linear function, variable coefficient of variation is large, the accuracy of each method are not ideal, law.