Scholarlink

Abstract <p>This paper presents a comprehensive analysis of training strategies for finite basis physics-informed neural networks (FBPINNs). We implement four distinct training approaches and evaluate them on six differential equations. Experimental results demonstrate that dependent simultaneous training achieves the highest accuracy but requires significant computational resources. We provide code on github and practical recommendations for selecting strategies based on problem complexity and resource constraints.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Analysis of Training Strategies for Finite Basis PINNs (FBPINNs)

  • 期刊论文

  • 发表日期：

    2025年12月1日

查看全文
  • P. G. Alimov,
  • V. I. Gorikhovskii

摘要

Abstract

This paper presents a comprehensive analysis of training strategies for finite basis physics-informed neural networks (FBPINNs). We implement four distinct training approaches and evaluate them on six differential equations. Experimental results demonstrate that dependent simultaneous training achieves the highest accuracy but requires significant computational resources. We provide code on github and practical recommendations for selecting strategies based on problem complexity and resource constraints.