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 2104.05337math.NAcs.NA
收藏

基于深度神经网络的各向异性椭圆型方程的渐近保持格式

Asymptotic preserving scheme for anisotropic elliptic equations with deep neural network

Long Li, Chang Yang

摘要:

本文对各向异性椭圆型方程提出了一种新的渐近保持格式。与以往的AP格式不同,实际的AP格式是基于二阶偏微分方程的一阶系统最小二乘,且关于各向异性强度是一致适定的.数值计算由深度神经网络(DNN)实现,其中采用最小二乘泛函作为损失函数来确定DNN的参数。数值结果表明,该方法易于实现,对各种二维和三维试验中的近似解或各向异性强度的识别具有鲁棒性。

英文摘要:

In this paper, a new asymptotic preserving (AP) scheme is proposed for the anisotropic elliptic equations. Different from previous AP schemes, the actual one is based on first-order system least-squares for second-order partial differential equations, and it is uniformly well-posed with respect to anisotropic strength. The numerical computation is realized by a deep neural network (DNN), where least-squares functionals are employed as loss functions to determine parameters of DNN. Numerical results show that the current AP scheme is easy for implementation and is robust to approximate solutions or to identify anisotropic strength in various 2D and 3D tests.

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。

↑