11.08.2013 Views

Untitled - Cdm.unimo.it

Untitled - Cdm.unimo.it

Untitled - Cdm.unimo.it

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Ordinary Differential Equations 193<br />

Proof - We start by considering ψ,φ ∈ P 0 n ⊂ H 1 0,w(I), for some n ≥ 2. Then, the<br />

first inequal<strong>it</strong>y is obtained by applying the Schwarz inequal<strong>it</strong>y (2.1.7) to the integral<br />

<br />

I (ψ′√ w)[(φw) ′ / √ w]dx. Recalling that (w ′ ) 2 /w ≤ w/(1 − x 2 ) 2 , we conclude using<br />

the results of lemma 8.6.1. The second inequal<strong>it</strong>y is proven in lemma 8.2.1. The third<br />

one comes from the Schwarz inequal<strong>it</strong>y and (5.7.5). The proof of (9.3.10), (9.3.11),<br />

(9.3.12) in H 1 0,w(I) requires a l<strong>it</strong>tle care. In short, one approximates φ ∈ H 1 0,w(I)<br />

by a sequence of polynomials φn ∈ P 0 n, n ∈ N, converging to φ in such a way that<br />

limn→+∞ φ − φn H 1 0,w (I) = 0. This also implies: limn→+∞ φn H 1 0,w (I) = φ H 1 0,w (I).<br />

The propos<strong>it</strong>ion follows by a lim<strong>it</strong> process which is standard in Lebesgue integration<br />

theory (see kolmogorov and fomin(1961)). The details are om<strong>it</strong>ted.<br />

Thus, (9.3.4) is a well-posed problem. In the special case when f ∈ C 0 ( Ī) ⊂ L2 w(I),<br />

the solution U of (9.3.4) is the classical strong solution satisfying (9.1.4).<br />

Generalizations can be examined w<strong>it</strong>h l<strong>it</strong>tle modification. For example, let us<br />

consider the boundary-value problem<br />

(9.3.13)<br />

⎧<br />

⎨<br />

−U ′′ + A1U ′ + A2U = f in I,<br />

⎩<br />

U(±1) = 0,<br />

where A1 : Ī → R and A2 : Ī → R are given bounded integrable functions. Here, we<br />

can replace (9.3.2) by the bilinear form<br />

(9.3.14) Bw(ψ,φ) =<br />

<br />

ψ<br />

I<br />

′ (φw) ′ dx +<br />

<br />

I<br />

[A1ψ ′ φ + A2ψφ]w dx, ∀ψ,φ ∈ X.<br />

One checks that, if the inequal<strong>it</strong>y A2w − 1<br />

2 (A1w) ′ ≥ 0 holds in I, then the hypotheses<br />

of theorem 9.3.1 are satisfied. Therefore, the variational formulation of (9.3.13) adm<strong>it</strong>s<br />

a unique weak solution.

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

Saved successfully!

Ooh no, something went wrong!