Adaptive Crank-Nicolson methods with dynamic finite-element spaces for parabolic problems (Q950572)

The authors construct discretizations of linear parabolic problems with the aid of Galerkin type methods, discontinuous in time and continuous in space. These procedures have the advantage that they are adaptive i.e., at each time level the refinement or coarsening of the triangulation can be modified according to the necessities. Important elements in this method are the a posteriori estimates. In order to explain their method the authors consider the following model problem: \[ \begin{alignedat}{2} u_{t}-\Delta u &=f \quad&&\text{in } \Omega \times (0,T), \\ u&=0 \quad&&\text{on }\partial\Omega \times (0,T),\\ u(\cdot,0)&=u_{0}\quad&&\text{in } \Omega,\end{alignedat} \] where \(\Omega\subset\mathbb{R}^{2}\), \(f\in L^{2}(0,T;L^{2}(\Omega))\), \(u_{0}\in L^{2}(\Omega)\). Next sections are: The approximate procedure; Crank-Nicolson procedure; Residual; Posing dual problems and representation formulas; A posteriori error estimator; Adaptive algorithm.