18.03.2014 Views

2 DGM for elliptic problems

2 DGM for elliptic problems

2 DGM for elliptic problems

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

18 2 Elliptic <strong>problems</strong><br />

a n h(u,v) = ∑<br />

K∈T h<br />

∫K<br />

−<br />

a i h(u,v) = ∑<br />

∑<br />

Γ ∈F ID<br />

h<br />

K∈T h<br />

∫K<br />

∇u · ∇v dx (2.45) ah_N<br />

∫<br />

Γ<br />

(n · 〈∇u〉 [v] − n · 〈∇u〉 [v]) dS,<br />

∇u · ∇v dx −<br />

∑<br />

Γ ∈F ID<br />

h<br />

∫<br />

Γ<br />

n · 〈∇u〉 [v]dS (2.46) ah_I<br />

and the linear <strong>for</strong>ms<br />

∫<br />

Fh(v) s = f v dx + ∑<br />

∫<br />

Fh n (v) =<br />

∫<br />

Fh(v) i =<br />

Ω<br />

Ω<br />

Γ ∈F N h<br />

f v dx + ∑<br />

Γ ∈F N h<br />

f v dx + ∑<br />

Ω<br />

Γ ∈F N h<br />

∫<br />

g N v dS + ∑ ∫<br />

n · ∇v u D dS, (2.47) Fh_S<br />

∫<br />

∫<br />

Γ<br />

Γ<br />

Γ<br />

Γ ∈F D h<br />

g N v dS − ∑<br />

∫<br />

Γ<br />

Γ ∈F D Γ<br />

h<br />

n · ∇v u D dS, (2.48) Fh_N<br />

g N v dS. (2.49) Fh_I<br />

Moreover, <strong>for</strong> u,v ∈ H 2 (Ω, T h ) let us define the bilinear <strong>for</strong>ms<br />

and the linear <strong>for</strong>ms<br />

B s h(u,v) = a s h(u,v), (2.50) A2.20a<br />

Bh(u,v) n = a n h(u,v), (2.51) A2.20b<br />

B s,σ<br />

h (u,v) = as h(u,v) + Jh σ (u,v), (2.52) A2.20c<br />

B n,σ<br />

h (u,v) = an h(u,v) + Jh σ (u,v), (2.53) A2.20d<br />

B i,σ<br />

h (u,v) = ai h(u,v) + Jh σ (u,v) (2.54) A2.20e<br />

def:2.1<br />

l s h(v) = F s h(v), (2.55) A2.21a<br />

l n h(v) = Fh n (v), (2.56) A2.21b<br />

l s,σ<br />

h (v) = F h(v) s + JD(v), σ (2.57) A2.21c<br />

l n,σ<br />

h (v) = F h n (v) + JD(v), σ (2.58) A2.21d<br />

l i,σ<br />

h (v) = F h(v) i + JD(v). σ (2.59) A2.21e<br />

Since S hp ⊂ H 2 (Ω, T h ), the <strong>for</strong>ms ( 2.50) A2.20a – ( 2.59) A2.21e make sense <strong>for</strong> u h ,v h ∈<br />

S hp . Consequently, we define five numerical schemes.<br />

Definition 2.8. A function u h ∈ S hp is called a DG approximate solution of<br />

problem ( 2.1) A0.1 – ( 2.3), A0.2b if it satisfies one of the following identities:<br />

i) B s h(u h ,v h ) = l s h(v h ) ∀v h ∈ S hp (2.60) A2.22a<br />

ii) B n h(u h ,v h ) = l n h(v h ) ∀v h ∈ S hp (2.61) A2.22b

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!