在一系列不同的极限状态函数条件下，对随机抽样法和拉丁超立方抽样法以及是否使用方差减缩技术进行了比较研究。
Comparative study on random sampling and Latin hypercube sampling with and without variance reduction techniques is carried out to a number of different limit state functions.
另一类变量与向量函数呈非线性关系。 对于后一类变量，用弃舍随机方法先给出位置初值，然后将问题化为线性最小二乘问题，直接解超定方程组。
The overdetermined equations can be solved directly by using the rejection method to give the initial value of the position first and then converting the problem into a linear—square one.
但是真实地震动的归一化自相关函数则有一较显著的随机噪声项，这个随机噪声的幅值为0。15左右，这就是本文中所说的超随机特性。
But the normalized auto-correlation of real seismic acceleration record comprises an obviously random noise with amplitude value of 0.15 or so, that is so called super random content.
但是真实地震动的归一化自相关函数则有一较显著的随机噪声项，这个随机噪声的幅值为0。15左右，这就是本文中所说的超随机特性。
But the normalized auto-correlation of real seismic acceleration record comprises an obviously random noise with amplitude value of 0.15 or so, that is so called super random content.
应用推荐