H AND H-P VERSION ERROR ESTIMATION AND ADAPTIVE PROCEDURES FROM THEORY TO PRACTICE

In this paper we introduce techniques that allow us to define a posterior error estimators via well-known recovery techniques. These allow us to construct a posteriori error estimators for relatively general problems. Further, we introduce new adaptive procedures that make use of these estimators and, in particular, describe an h-p procedure that is simple to implement and that, as numerical experiments have shown, attains an accelerated rate of convergence expected from the h-p version.