English

An explicit and symmetric exponential wave integrator for the nonlinear Schr\"{o}dinger equation with low regularity potential and nonlinearity

Numerical Analysis 2025-04-29 v2 Numerical Analysis

Abstract

We propose and analyze a novel symmetric Gautschi-type exponential wave integrator (sEWI) for the nonlinear Schr\"odinger equation (NLSE) with low regularity potential and typical power-type nonlinearity of the form ψ2σψ |\psi|^{2\sigma}\psi with ψ \psi being the wave function and σ>0 \sigma > 0 being the exponent of the nonlinearity. The sEWI is explicit and stable under a time step size restriction independent of the mesh size. We rigorously establish error estimates of the sEWI under various regularity assumptions on potential and nonlinearity. For ``good" potential and nonlinearity (H2H^2-potential and σ1\sigma \geq 1), we establish an optimal second-order error bound in the L2L^2-norm. For low regularity potential and nonlinearity (LL^\infty-potential and σ>0\sigma > 0), we obtain a first-order L2L^2-norm error bound accompanied with a uniform H2H^2-norm bound of the numerical solution. Moreover, adopting a new technique of \textit{regularity compensation oscillation} (RCO) to analyze error cancellation, for some non-resonant time steps, the optimal second-order L2L^2-norm error bound is proved under a weaker assumption on the nonlinearity: σ1/2\sigma \geq 1/2. For all the cases, we also present corresponding fractional order error bounds in the H1H^1-norm, which is the natural norm in terms of energy. Extensive numerical results are reported to confirm our error estimates and to demonstrate the superiority of the sEWI, including much weaker regularity requirements on potential and nonlinearity, and excellent long-time behavior with near-conservation of mass and energy.

Keywords

Cite

@article{arxiv.2310.20181,
  title  = {An explicit and symmetric exponential wave integrator for the nonlinear Schr\"{o}dinger equation with low regularity potential and nonlinearity},
  author = {Weizhu Bao and Chushan Wang},
  journal= {arXiv preprint arXiv:2310.20181},
  year   = {2025}
}

Comments

28 pages, 6 figures