Unformatted text preview: kp+1:
This estimate does not have an optimal rate since the interpolation error (7.2.10a) is converging as O(hp+1). Getting the correct rate for an L2 error estimate is more complicated
than it is in H 1. The proof is divided into two parts. 7.2. Convergence and Optimality 9 Lemma 7.2.1. (Aubin-Nitsche) Under the assumptions of Theorem 7.2.3, let (x y) 2
H01 be the solution of the \dual problem" A(v ) = (v e)
8v 2 H0 u;U
e = ku ; U k :
Let ; 2 S0 be an interpolant of , then ku ; U k0 ku ; U k1 k ; ;k1 : (7.2.12c) Proof. Set V = ; in (7.2.6) to obtain A(; u ; U ) = 0: (7.2.13) Take the L2 inner product of (7.2.12b) with u ; U to obtain
ku ; U k0 = (e u ; U ): Setting v = u ; U in (7.2.12a) and using the above relation yields
ku ; U k0 = A(u ; U ): Using (7.2.13) ku ; U k0 = A(u ; U ; ;):
Now use the continuity of A(v u) in H 1 ((7.2.1) with s = 1) to obtain (7.2.12c). Since we have an estimate for ku ; U k1 , estimating ku ; U k0 by (7.2.12c) requires
an estimate o...
View Full Document
This document was uploaded on 03/16/2014 for the course CSCI 6860 at Rensselaer Polytechnic Institute.
- Spring '14