Exponential Low-Regularity Parareal Algorithms for Nonlinear Schrödinger Equations (Q7270446)

The paper designs linearly convergent parareal algorithms with exponential low-regularity coarse propagators for the nonlinear Schrödinger equation, which are provably accurate and efficient even when solutions have limited regularity. These methods achieve optimal approximation orders without derivative loss and outperform variants using classical coarse propagators in numerical experiments on one-dimensional quadratic, cubic, and quintic NLS equations.