Effective condition number for the finite element method using local mesh refinements (Q1030666)

This paper is a continued study of the effective condition number in [\textit{Z. C. Li, C. S. Chien} and \textit{H. T. Huang}, J. Comput. Appl. Math. 198, No. 1, 208--235 (2007; Zbl 1104.65043); \textit{Z. C. Li} and \textit{H. T. Huang}, Number. Linear Algebra Appl. 15, 575--594 (2008)] for stability analysis. To approximate Poisson's equation with singularity by the finite element method (FEM), the adaptive mesh refinements are an important technique, by which the FEM solutions with optimal convergence rates can be obtained. However, the traditional condition number given by \(\text{Cond}={\mathcal O}(h^{-2}_{\min})\) in [\textit{G. Strang} and \textit{G. J. Fix}, An analysis of the finite element method. Englewood Cliffs, N.J.: Prentice-Hall (1973; Zbl 0356.65096)], where \(h_{\min}\) is the minimal length of elements, is huge. In this paper, the bounds of the simplified effective condition number \(\text{Cond}_{EE}\) are derived as \({\mathcal O}(1), {\mathcal O}(h^{-1.5}), {\mathcal O}(h^{-0.5})\), where \(h (<1)\) is the maximal length of elements. It is clear that \(\text{Cond}_{EE}\) is smaller than Cond. Numerical experiments are carried out, to verify the stability analysis. Small effective condition numbers explain well the satisfactory FEM solutions obtained. This paper improves previous results, and provides a stability justification for the adaptive mesh refinements used in FEM.