On \(L^2\)-error estimate for nonsymmetric interior penalty Galerkin approximation to linear elliptic problems with nonhomogeneous Dirichlet data (Q1019767)

The authors discuss \(L^2\)-error estimates for the nonsymmetric interior penalty Galerkin approximation applied to the Poison equation \(-\Delta u=f\) in \(\Omega \) with smooth nonhomogeneous data \(u=g\) on \(\partial G\), employing super-penalty. Using piecewise polynomials of degree less than or equal to \(r,\) they obtain a new estimate which is of order \(\left( h/r\right) ^{r+1/2}\) when \(g\in H^{r+1/2}\left( \partial \Omega \right) \) and is optimal, i.e. of order \(\left( h/r\right) ^{r+1}\) when \(g\in H^{r+1}\left( \partial \Omega \right) \), where \(h\) denotes the mesh size. Numerical experiments illustrate the theoretical results.