{"id":"arxiv/2606.09676","url":"https://arxiv.org/abs/2606.09676","tags":["math.NA"],"score":0,"title":"Boundary-Layer-Induced Failure of Standard Physics-Informed Neural Networks: A Legendre Wavelet Collocation Benchmark for Singularly Perturbed Transport Problems","domain":"math.NA","source":"arxiv","authors":["Suvendu Nayak","Arun Kumar Gupta"],"reviews":0,"subject":"math.NA","summary":"Boundary layers provide a demanding test for numerical solvers because the solution may remain almost constant over most of the domain while changing rapidly in a narrow region near the boundary. This paper studies a singularly perturbed one-dimensional transport boundary-value problem with increasing Peclet number $(\\mathrm{Pe})$. A local Legendre wavelet collocation method (LWM) is compared with a standard soft-boundary physics-informed neural network (PINN) for this benchmark. The wavelet approximation uses locally supported Legendre polynomial basis functions and converts the problem…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-06-08","reactions":{}}