A posteriori error estimation and basis adaptivity for reduced-basis approximation of nonaffine-parametrized linear elliptic partial differential equations (Q2464971)

For parametrized partial differential equations (PDE)s the reduced-basis solutions (obtained via a Galerkin finite element approximation) of linear elliptic equations with nonaffine parameter dependence resides on a low-dimensional manifold. The author provides a posteriori error estimation and basis adaptivity for the reduced basis approximation of linear elliptic PDEs. The new contributions are the a priori convergence analysis, the derivation of a posteriori error estimators and the optimal construction of hierarchical reduced-basis spaces based on the greedy sampling procedure and a proper orthogonal decomposition or the Karhunen-Loéve approach.