Adaptive least-squares finite element methods: Guaranteed upper bounds and convergence in L 2 norm of the dual variable

JaEun Ku
DOI: https://doi.org/10.1016/j.camwa.2024.03.002
IF: 3.218
2024-03-16
Computers & Mathematics with Applications
Abstract:We consider adaptive least-squares finite element methods. First, we develop a guaranteed upper bound for the dual error in the L2 norm, and this can be used as a stopping criterion for the adaptive procedures. Secondly, based on the a posteriori error estimates for the dual variable, we develop an error indicator that identifies the local area to refine, and establish the convergence of the adaptive procedures based on the Dörfler's marking strategy. Our convergence analysis is valid for the entire range of the bulk parameter 0<Θ≤1 and it shows the effect of bulk parameter and reduction factor of elements on the convergence rate. Confirming numerical experiments are provided.
mathematics, applied
What problem does this paper attempt to address?