Extension of a Post-Processing Technique for the Discontinuous
Galerkin Method

 

Presented by:

 

Jennifer K. Ryan

Householder Postdoctoral Fellow

Computer Science and Mathematics Division
Oak Ridge National Laboratory

 

April 16, 200410:30 amSuite 150, Mesa Laboratory

 

 

 

 

Abstract:

 

Cockburn, Luskin, Shu, and Suli previously introduced a post-processing technique for the discontinuous Galerkin method that allows improvement in accuracy from order k+1 to order 2k+1 for linear hyperbolic equations.  This technique is based on negative order norm estimates.  A uniform mesh assumption allows for simple implementation via small matrix-vector multiplications.  In this talk, we present extensions of this improvement in accuracy to include superconvergence of the derivatives, variable coefficient hyperbolic equations, one-sided post-processing, and post-processing over a smoothly varying mesh.