arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 0704.2094math.NAcs.NA
收藏

泊松方程的前置小波解

Prewavelet Solution to Poisson Equations

University of Georgia(佐治亚大学)

Ming-Jun Lai, Haipeng Liu

AI总结:

提出一种前置小波方法,通过添加前置小波子解来保留原始有限元解,从而高效地提高泊松方程数值解的精度。

摘要:

有限元法是求解偏微分方程的有力数值方法之一。通常,如果基于底层区域三角剖分的泊松方程有限元解不够精确,人们会丢弃该解,然后均匀细化三角剖分并在细化后的三角剖分上计算新的有限元解。丢弃原始有限元解是一种浪费。我们提出了一种前置小波方法,通过添加前置小波子解来保存原始解,从而获得细化后的有限元解。为了提高泊松方程数值解的精度,我们可以不断添加前置小波子解。我们的前置小波在$H^1$范数下正交,并且除了矩形域中的一个全局支撑基函数外,它们具有紧支撑。我们在MATLAB中实现了这些前置小波基函数,并将其用于具有Dirichlet边界条件的泊松方程的数值求解。数值模拟表明,我们的前置小波解比标准有限元方法高效得多。

英文摘要:

Finite element method is one of powerful numerical methods to solve PDE. Usually, if a finite element solution to a Poisson equation based on a triangulation of the underlying domain is not accurate enough, one will discard the solution and then refine the triangulation uniformly and compute a new finite element solution over the refined triangulation. It is wasteful to discard the original finite element solution. We propose a prewavelet method to save the original solution by adding a prewavelet subsolution to obtain the refined level finite element solution. To increase the accuracy of numerical solution to Poisson equations, we can keep adding prewavelet subsolutions. Our prewavelets are orthogonal in the $H^1$ norm and they are compactly supported except for one globally supported basis function in a rectangular domain. We have implemented these prewavelet basis functions in MATLAB and used them for numerical solution of Poisson equation with Dirichlet boundary conditions. Numerical simulation demonstrates that our prewavelet solution is much more efficient than the standard finite element method.

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。

↑