[Home]   [  News]   [  Events]   [  People]   [  Research]   [  Education]   [Visitor Info]   [UCSD Only]   [Admin]
Home > Events > CCoM > Abstract
Search this site:


Directors:
Randolph E. Bank
Philip E. Gill
Michael Holst

Administrative Contact:
Terry Le

Office: AP&M 7431
Phone: (858)534-9813
Fax: (858)534-5273
E-mail: tele@ucsd.edu
Least-squares finite element methods for the Poisson equation and their connection to the Dirichlet and Kelvin principles

Max Gunzburger
Mathematics and School of Computational Science, Florida State University

Abstract:

Least-squares finite element methods for first-order formulations ofthe Poisson equation are not subject to the inf-sup condition andlead to stable solutions even when all variables are approximated byequal-order continuous finite element spaces. For such elements, onecan also prove optimal convergence in the ``energy' norm for allvariables and optimal L^2 convergence for the scalar variable.However, showing optimal L^2 convergence for the flux has proven tobe impossible without adding the redundant curl equation to the first-order system of partial differential equations. In fact, numericalevidence strongly suggests that nodal continuous flux approximationsdo not posses optimal L^2 accuracy. In this talk, we show thatoptimal L^2 error rates for the flux can be achieved without the curlconstraint, provided that one uses the div-conforming family ofBrezzi-Douglas-Marini or Brezzi-Douglas-Duran-Fortin elements. Then,we proceed to reveal an interesting connection between a least-squares finite element method involving div-conforming fluxapproximations and mixed finite element methods based on theclassical Dirichlet and Kelvin principles. We show that such least-squares finite element methods can be obtained by approximating,through an L^2 projection, the Hodge operator that connects theKelvin and Dirichlet principles. Our principal conclusion is thatwhen implemented in this way, a least-squares finite element methodcombines the best computational properties of mixed-Galerkin finiteelement methods based on each of the classical principles.

Thursday, May 17, 2007
11:00AM AP&M 5402