Monotone iterates with quadratic convergence rate for solving semilinear parabolic problems (Q2896254)

The author considers a \(d\)-dimensional semilinear parabolic problem along with Dirichlet boundary data, discretized so as to obtain an \(M\)-matrix (to allow for the application of Samarskij's discrete maximum principle), and develops a method of monotone inclusion on the different time levels, providing a constructive way to get upper and lower solutions to start the monotone iteration on any time level. He proves monotone, second order convergence, existence and uniqueness of the solution of the discrete equations, discusses stopping criteria and shows a series of numerical results for two-dimensional, singularly perturbed convection-diffusion-reaction equations, using a Shishkin mesh, solving the discrete equations by the generalized minimal residual (GMRES) method. His results compare advantageously to own earlier ones, [J. Numer. Math. 14, No. 4, 247--266 (2006; Zbl 1125.65078)].NEWLINENEWLINE It should be remarked that in the formulation of the iterative method, there appears an undefined function, and no mention of earlier work of the Dresden school of monotone inclusion methods (J. W. Schmidt etc.) is made.