一般阻尼随机微分方程的拟局部振荡算法

A Quasi-local Vibration Algorithm for Damped Stochastic Differential Equations

  • 摘要: 提出了一种模拟随机微分方程的拟局部振荡算法,即利用算符劈裂方法和势函数的泰勒展开,对噪声作用下耗散粒子的时间演化算符进行分解,得到了对应涨落行为的扩散算符和对应确定轨迹的漂移算符.其中局部简谐势场的涨落过程可获得解析解,而剩余的确定项则利用简单的Euler算法积分.应用到几个算例并与常用的两种算法相比较,结果表明:本算法随时间步长最稳定,可使用较大的时间步长.

     

    Abstract: A quasi-local vibration algorithm is proposed to simulate the damped stochastic differential equations.Using the potential Taylor expansion and the operator splitting technique,the time evolution operator of a dissipative particle is factorized into two parts: an operator which describes the deterministic path of the system and a diffusion operator which incorporates fluctuation away from path.The differential equation corresponding to the fluctuation process is solved analytically, and the deterministic trajectory equation is calculated with the Euler algorithm.Compared with other algorithms, it is shown that the calculated results by the present algorithm are more stable and converge to the correct value when the time step is large.

     

/

返回文章
返回