随机扩散模型一种新的密度函数统计方法

张蕾; 王立良; 高远

doi:doi:10.11884/hplpb201729.170268

强激光与粒子束, 2017, 29 (12): 126009, 网络出版: 2017-12-25

随机扩散模型一种新的密度函数统计方法

A statistical method for the density function of the stochastic differential model function

论文大纲

张蕾王立良高远

作者单位

杭州电子科技大学信息工程学院, 杭州 310018

随机扩散方程扩散输运方程拉格朗日模型核函数法 stochastic diffusion equation advection-diffusion transport equation Lagrangian model kernel function method

摘要

引入核函数法对随机扩散方程（SDE）样本的密度分布进行统计, 希望用核函数来减少统计涨落。由于SDE样本的密度随时间发展, 越来越稀疏, 所以核函数也应该越来越大, 也就是说核函数应该随时间在变化。通过一个瞬时释放的二维扩散问题（具有解析解）, 从定性和定量两个角度比较了变带宽核函数法和传统统计方法在密度分布统计中的性能差别, 论述了变带宽核函数法的优缺点, 变带宽核函数法在牺牲部分峰值的前提下可以很好地解决SDE样本密度分布统计涨落问题, 在工程应用中值得推广。

Abstract

In this paper, the density distribution of stochastic diffusion equation (SDE) is calculated by the kernel function method, and the statistical fluctuation is reduced by the kernel function. As the density of SDE over time is sparser and sparser, the kernel function is bigger and bigger, i.e., the kernel function is changing over time, and the rule is given by a method called the variable bandwidth kernel function method. Through an instantaneous release of a two-dimensional diffusion problem (with the analytical solution), from the perspective of both qualitative and quantitative comparisons of the variable bandwidth kernel function method and the traditional statistical method in the performance of density distribution statistical difference, this paper discusses the advantages and disadvantages of variable bandwidth kernel function method. The variable bandwidth kernel function method which sacrifices part of the peak can be very good to solve the problem of SDE density distribution statistical fluctuation, and it is worth promoting in engineering applications.

参考文献

[1] Monti P, Leuzzi G. A closure to derive a three-dimensional well-mixed trajectory model for non-Gaussian, inhomogeneous turbulence［J］. Boundary-Layer Meteorol, 1996, 80(4): 311-331.

[2] Leuzzi G, Purini R. The effect of a promontory on the passive tracers advected by a coastal current: A full three-dimensional numerical study［J］. IL Nuovo Cimento C, 2005, 28(28): 61-74.

[3] Riddle A M. The specification of mixing in random walk models for dispersion in the sea［J］. Continental Shelf Research, 1998, 18(2/4): 441-456.

[4] Silverman B W. Density estimation for statistics and data analysis［M］. London: Chapman and Hall, 1986.

[5] Darya S A, Heemink E W. Lagrangian modelling of multi-dimensional advection-diffusion with space-varying diffusivities: Theory and idealized test cases［J］. Ocean Dynamics, 2007, 57(3): 189-203.

[6] 阳林峰,杨宏伟,骆志平,等.利用核函数提高随机游走模式中污染物浓度计算的效率［J］.原子能科学技术, 2015, 49(3): 485-490.(Yang Linfeng, Yang Hongwei, Luo Zhiping, et al. Using kernal function to improve efficiency of contaminated concentration calculation in random walk model. Atomic Energy Science and Technology, 2015, 49(3): 485-490)

[7] Wade E H. Simplified approach to particle tracking methods for contaminant transport［J］. Hydraulic Engineering, 1997, 123(12): 1157-1160.

[8] Cristianini N, Shawe-Taylor J. Kernel methods for pattern recognition［M］. Cambridge: Cambridge University Press, 2004.

张蕾, 王立良, 高远. 随机扩散模型一种新的密度函数统计方法[J]. 强激光与粒子束, 2017, 29(12): 126009. Zhang Lei, Wang Liliang, Gao Yuan. A statistical method for the density function of the stochastic differential model function[J]. High Power Laser and Particle Beams, 2017, 29(12): 126009.

随机扩散模型一种新的密度函数统计方法

关于本站 Cookie 的使用提示

全站搜索

随机扩散模型一种新的密度函数统计方法

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索