{"id":"arxiv/2603.15331","url":"https://arxiv.org/abs/2603.15331","tags":["math.NA","cs.LG"],"score":0,"title":"A scaled TW-PINN: A physics-informed neural network for traveling wave solutions of reaction-diffusion equations with general coefficients","domain":"math.NA","source":"arxiv","authors":["Seungwan Han","Kwanghyuk Park","Jiaxi Gu","Jae-Hun Jung"],"reviews":0,"subject":"math.NA","summary":"We propose an efficient and generalizable physics-informed neural network (PINN) framework for computing traveling wave solutions of $n$-dimensional reaction-diffusion equations with various reaction and diffusion coefficients. By applying a scaling transformation with the traveling wave form, the original problem is reduced to a one-dimensional scaled reaction-diffusion equation with unit reaction and diffusion coefficients. This reduction leads to the proposed framework, termed scaled TW-PINN, in which a single PINN solver trained on the scaled equation is reused for different coefficient…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-16","reactions":{}}