Id: arxiv/2607.23143 Url: https://arxiv.org/abs/2607.23143 Tags: math.NA, physics.flu-dyn Score: 0 Title: Physics-informed token transformer methodology for nonlinear balance laws. I. Schwarzschild--Burgers fluid flows Domain: math.NA Source: arxiv Authors: Philippe G. LeFloch, Shuyang Xiang Reviews: 0 Subject: math.NA Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-07-25 Page: https://delphi.science/r/arxiv/2607.23143 We introduce a Physics-Informed Token Transformer (PITT) methodology for nonlinear hyperbolic balance laws in one space dimension, using piecewise steady-state profiles for the representation of approximate weak solutions. The method combines symbolic equation tokenization, a Fourier neural operator encoder, an explicit Rankine--Hugoniot law for shock motion, and a learned correction term. For clarity, we present it here for the relativistic Schwarzschild--Burgers equation, a scalar model for spherically symmetric fluid flows on a Schwarzschild background. For this model the steady-state…