PDE-independent adaptive \(hp\)-FEM based on hierarchic extension of finite element spaces (Q961529)

The authors present a novel approach to automatic adaptivity in higher-order finite element methods (\(hp\)-FEM). The goal is to develop adaptive \(hp\)-FEM algorithms with controlled accuracy in both space and time. This approach is free of analytical error estimates. A posteriori error estimation is done computationally via hierarchic extension of finite element spaces. This is an analogy with embedded higher-order methods for ODE. The adaptivity process yields a sequence of embedded stiffness matrices which are solved efficiently using a simple combined direct-iterative algorithm. The methodology works well for standard low-order FEM and for \(hp\)-FEM. Numerical examples are presented. The preliminary results indicate that the iterative method converges fast both for symmetric positive definite matrices and for strongly ill-conditioned indefinite matrices. The method needs to be further studied and optimized. Convergence of the iterative process needs also to be analysed.