11.08.2013 Views

Untitled - Cdm.unimo.it

Untitled - Cdm.unimo.it

Untitled - Cdm.unimo.it

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.

228 Polynomial Approximation of Differential Equations<br />

A special case worth mentioning is I ≡ R in (10.2.1) and (10.2.2). Of course,<br />

we no longer have boundary cond<strong>it</strong>ions. These are supplanted by the cond<strong>it</strong>ion that<br />

U tends to zero at infin<strong>it</strong>y w<strong>it</strong>h a specified minimum rate. We derive a collocation<br />

type approximation in this case. According to escobedo and kavian(1987), first set<br />

V (x,t) := U( √ 4ζxe2ζt ,e4ζt − 1), x ∈ R, t ∈]0, ˆ T], where ˆ T := 1<br />

4ζln(T + 1). Then, <strong>it</strong><br />

is clear that V satisfies the equation<br />

(10.2.20)<br />

∂V<br />

(x,t) = ζ<br />

∂t<br />

2 ∂ V<br />

(x,t) + 2x∂V<br />

∂x2 ∂x (x,t)<br />

<br />

, ∀x ∈ R, ∀t ∈]0, ˆ T],<br />

(10.2.21) V (x,0) = U( 4ζx,0) = U0( 4ζx), ∀x ∈ R.<br />

Assuming that the solution V decays as fast as e −x2<br />

at infin<strong>it</strong>y, we approximate by<br />

Herm<strong>it</strong>e functions (see section 6.7). Let ξ (n)<br />

i , 1 ≤ i ≤ n, be the zeroes of Hn. For any<br />

t ∈]0, ˆ T], we want to find Pn(·,t) ∈ Sn−1 such that<br />

(10.2.22)<br />

∂Pn<br />

∂t (ξ(n)<br />

i ,t) = ζ<br />

2 ∂ Pn<br />

(ξ(n)<br />

∂x2 i ,t) + 2ξ (n)<br />

i<br />

(10.2.23) Pn(ξ (n)<br />

i ,0) = U0( 4ζξ (n)<br />

i ), 1 ≤ i ≤ n.<br />

∂Pn<br />

∂x (ξ(n)<br />

<br />

i ,t) , 1 ≤ i ≤ n,<br />

We can formulate the same problem in the space of polynomials by setting pn := Pne x2<br />

(see section 7.4). This is equivalent to a system of ordinary differential equations. The<br />

corresponding matrix has non-pos<strong>it</strong>ive eigenvalues, as pointed out in section 8.2, which<br />

is very important for proving convergence of Pn to V for n → +∞. Actually,<br />

recalling formula (3.4.1), the proof follows the same arguments as the case when I is<br />

bounded. The theory is presented in funaro and kavian(1988). Finally, we can get<br />

an approximation of the unknown U w<strong>it</strong>h the change of variables: x → x[4ζ(t+1)] −1/2 ,<br />

t → 1<br />

4ζln(t + 1).<br />

We remark that the direct approximation of the unknown U by the Herm<strong>it</strong>e-<br />

collocation method applied to equation (10.2.1) w<strong>it</strong>h I ≡ R, is unstable. This is due<br />

to the fact that the quant<strong>it</strong>y <br />

R ψ′ (ψv) ′ dx is not in general pos<strong>it</strong>ive for ψ ∈ H 1 v(R),<br />

where v(x) := e x2<br />

, x ∈ R.

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

Saved successfully!

Ooh no, something went wrong!