Numerical analysis of fully discretized Crank-Nicolson scheme for fractional-in-space Allen-Cahn equations (Q2412739)

The authors study the numerical approximation of the fractional-in-space Allen-Cahn (FiSAC) equation where the nonlinear term \(f (u)\) is taken as the polynomial double-well potential. In order to verify whether the maximum principle still holds for the numerical solutions of the FiSAC equation with high-order temporal discretizations, the second-order Crank-Nicolson (CN) scheme in time and second-order central difference approximation in space is employed. The resulting nonlinear system is solved by a nonlinear iteration algorithm which can reduce the total computation cost significantly. The iterations successfully avoid the requirement of inverting a dense matrix and can reduce the total computation cost to \(\mathcal{O}(N \log N)\) with \(N\) unknown freedoms by using the discrete Fourier transformation. The predominant contribution made by the article is the proof that the fully discretized scheme satisfies the numerical maximum principle with a reasonable time step constraint. This discrete maximum principle allows for the establishment of the nonlinear energy stability and the error estimates in maximum norm. The numerical scheme is shown to be second order accurate in both space and time, i.e., the error is of the rate \(\mathcal{O}(\Delta x^2 + \Delta t^2)\). Numerical examples are provided. It is stated that all results and analyses can be extended to standard Allen-Cahn equations trivially.