An introduction to the a posteriori error analysis of parabolic partial differential equations
Iain Smears
TL;DR
This work surveys the core ideas of a posteriori error analysis for parabolic PDEs, focusing on how the choice of error norm and the reconstruction of time-discrete solutions shape estimator effectiveness. It develops inf-sup stability frameworks for multiple norms, derives residual-based and equilibrated-flux estimators, and introduces time- and space-discretization reconstructions that underpin sharp a posteriori bounds. Key contributions include an energy-norm perspective with a hypercircle theorem, a detailed semi-discrete and fully discrete analysis using equilibrated fluxes, and a discussion of data-oscillation effects and local efficiency limits, informing adaptive strategies for time-dependent problems. The results illuminate how to design robust, locally efficient, and norm-adapted estimators that support practical adaptive simulations of heat-like parabolic problems.
Abstract
This article provides a brief introduction to the a posteriori error analysis of parabolic partial differential equations, with an emphasis on challenges distinct from those of steady-state problems. Using the heat equation as a model problem, we examine the crucial influence of the choice of error norm, as well as the choice of notion of reconstruction of the discrete solution, on the analytical properties of the resulting estimators, especially in terms of the efficiency of the estimators.