A posteriori error estimation of the p-Laplace problem

Abstract

Two a posteriori error estimates are discussed for the p-Laplace problem. Up to errors in their numerical computation, they provide a guaranteed upper bound for the W1,p-seminorm and a weighted W1,2-seminorm of u – uh. The first, sharper a posteriori estimate is based on the numerical solution of local interface problems. The standard residual-based error estimate is addressed with emphasis on involved constants, determined as local eigenvalues. Numerical examples that illustrate the performance of these estimators can be found in [3].