随机Landau-Lifshitz方程的稳定快速半隐式积分
Stable and fast semi-implicit integration of the stochastic Landau-Lifshitz equation
AI总结:
提出具有角动量守恒特性的半隐式数值方法,用于随机Landau-Lifshitz方程的积分,在一维海森堡链上测试,相比显式方法稳定性更好、步长更大,且比中点隐式方法计算成本更低。
摘要:
我们提出了新的半隐式数值方法,用于积分具有内置角动量守恒的随机Landau-Lifshitz方程。所提出的积分器的性能在一维海森堡链上进行了测试。对于该系统,我们的方案显示出更好的稳定性,并允许使用比标准显式方法大得多的时间步长。同时,这些半隐式方案也具有与标准中点隐式方法相当的精度,且计算成本低得多。这些结果对于原子尺度自旋动力学模拟以及超越宏观自旋近似的自旋动力学研究至关重要。
英文摘要:
We propose new semi-implicit numerical methods for the integration of the stochastic Landau-Lifshitz equation with built-in angular momentum conservation. The performance of the proposed integrators is tested on the 1D Heisenberg chain. For this system, our schemes show better stability properties and allow us to use considerably larger time steps than standard explicit methods. At the same time, these semi-implicit schemes are also of comparable accuracy to and computationally much cheaper than the standard midpoint implicit method. The results are of key importance for atomistic spin dynamics simulations and the study of spin dynamics beyond the macro spin approximation.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。