arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.08362math.NAcs.NAphysics.comp-ph
收藏

深度 Ritz-有限元方法:用有限元训练的神经网络方法

Deep Ritz-Finite Element methods: Neural Network Methods trained with Finite Elements

Georgios Grekas, Charalambos G. Makridakis

AI总结:

本文提出深度 Ritz-有限元方法,通过有限元空间计算损失积分来训练神经网络,用于求解低维椭圆型偏微分方程,兼具稳定性、收敛性及高效鲁棒性。

摘要:

虽然神经网络方法的许多关注都集中在高维偏微分方程问题上,但在本工作中,我们考虑的是针对定义在区域 Ω⊂ℝ^d(d=1,2,3)上的椭圆型问题、并与更标准的有限元相结合而设计的方法。我们建议通过训练将有限元与神经网络逼近联系起来,即利用有限元空间来计算损失泛函中出现的积分。这种方法保留了经典神经网络方法求解偏微分方程的简洁性,同时利用成熟的有限元工具(和软件)来计算所涉及的积分,从而在效率和精度上获得提升。我们证明了所提方法是稳定的,并且进一步建立了所得逼近收敛到偏微分方程解的理论结果。文中还给出了数值结果,以表明所提算法的高效性和鲁棒性。

英文摘要:

While much attention of neural network methods is devoted to high-dimensional PDE problems, in this work we consider methods designed to work for elliptic problems on domains $Ω\subset \mathbb{R} ^d, $ $d=1,2,3$ in association with more standard finite elements. We suggest to connect finite elements and neural network approximations through training, i.e., using finite element spaces to compute the integrals appearing in the loss functionals. This approach, retains the simplicity of classical neural network methods for PDEs, uses well established finite element tools (and software) to compute the integrals involved and it gains in efficiency and accuracy. We demonstrate that the proposed methods are stable and furthermore, we establish that the resulting approximations converge to the solutions of the PDE. Numerical results indicating the efficiency and robustness of the proposed algorithms are presented.

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。

↑