According to the wavelet deniosing method, a new algorithm for generating standard normal random numbers is proposed in this paper. For the standard normal random number generated by randn function, a comparative study is done to discuss the influence of different threshold rules on the mean and variance of random number, the influence of different decomposition levels on random number. Then the correlation among the components of high-dimensional random number is discussed in different space scales. For 1000 groups of normal random number, the distributions of p value of J-B test, mean, variance and correlation are shown by their boxplot. WMC method is presented and applied in numerical integration. For 1000 groups of approximation values computed by WMC method, the mean and variance are given for discussing its accuracy and stability by the boxplot. Finally, an example is given for numerical simulation of financial model.

根据小波去噪方法，本文提出了一种新的标准正态随机数生成算法。针对randn函数生成的标准正态随机数，通过对比研究，讨论了不同阈值规则对随机数均值和方差的影响，不同分解级别对随机数的影响。然后讨论了不同空间尺度下高维随机数各分量之间的相关性。对于 1000 组正态随机数，p的分布JB 检验的值、均值、方差和相关性由它们的箱线图显示。提出了 WMC 方法并将其应用于数值积分。对于 WMC 方法计算的 1000 组近似值，给出均值和方差，以通过箱线图讨论其准确性和稳定性。最后给出了一个金融模型数值模拟的例子。