随机过程的概率密度函数估计  

Probability density function estimation of stochastic processes

作　　者：李杰[1] 陈建兵[1] 张琳琳[1] 

机构地区：[1]同济大学,上海200092

出　　处：《应用力学学报》2010年第3期481-485,共5页Chinese Journal of Applied Mechanics

基　　金：国家自然科学基金创新研究群体科学基金(50621062);国家自然科学基金(10872148);863计划项目(2008AA05Z413)

摘　　要：从概率密度演化理论的基本思想出发,发展了随机过程一维概率密度函数估计的新方法。以独立获取的各随机过程样本作为随机过程的代表性时程,通过求解广义密度演化方程,获得了随机过程的一维概率密度函数及其均值与标准差过程。以脉动风速随机过程的统计为例,进行了风速时程的概率密度函数估计,为认识随机过程的概率结构提供了新的可能。In estimation of statistics of stochastic processes, usually either the correlation functions or the power spectral density functions are statistically computed to capture the second order information of the processes. In the present paper, started from the probability density evolution method, a new approach to estimate the one-dimensional probability density function (PDF) of stochastic processes is developed. The independently observed samples of a stochastic process are taken as the representative processes, the generalized probability density evolution equation is then solved to attain the one-dimensional PDF and simultaneously time histories of the mean and the standard deviation. Taking the statistics of wind velocity fluctuations as an example, theproposed approach is illustrated. The approach provides new possibility of recognizing the probabilistic structure of stochastic processes.

关 键 词：随机过程 概率密度函数 密度演化方法 差分方法 脉动风速 

分 类 号：C302[社会学]