Third-order Accurate, Large Time-Stepping and Maximum-Principle-preserving Schemes for the Allen-Cahn Equation

Hong Zhang,Xu Qian,Songhe Song
DOI: https://doi.org/10.1007/s11075-023-01606-w
IF: 2.37
2024-01-01
Numerical Algorithms
Abstract:We present and evaluate several explicit, large time-stepping algorithms for the Allen-Cahn equation. Our approach incorporates a stabilization technique and uses Taylor series approximations for exponential functions to develop a family of up to third-order parametric Runge–Kutta schemes that maintain fixed-points and maximum principle for any time step τ > 0 . We also introduce a new relaxation technique that eliminates time delay caused by stabilization. To further decrease the stabilization parameter, we utilize an integrating factor with respect to the stiff linear operator and develop a parametric relaxation integrating factor Runge–Kutta (pRIFRK) framework. Compared to existing maximum-principle-preserving (MPP) schemes, the proposed parametric relaxation approaches are free from limiters, cut-off post-processing, exponential decay, or time delay. Linear stability analysis determines that the parametric approaches are A -stable when appropriate parameters are used. In addition, we provide error estimates in the l^∞ -norm with the help of the MPP property. We demonstrate the high-order temporal accuracy, maximum-principle-preservation, energy stability, and delay-free properties of the proposed schemes through a set of experiments on 1D, 2D, and 3D problems.
What problem does this paper attempt to address?