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.

Numerical Integration 47<br />

This means that, by knowing the values p(ξ (n)<br />

j ), 1 ≤ j ≤ n, we can for instance recover<br />

the point values of Iw,np 2 . These are p 2 (ξ (n)<br />

j ), 1 ≤ j ≤ n. Hence, the value in x of<br />

the interpolant of p 2 is n<br />

j=1 p2 (ξ (n)<br />

j )l (n)<br />

j (x).<br />

This trivial remark is a key point in the treatment of nonlinear terms in the approxi-<br />

mation of differential equations (see section 9.8).<br />

d (α,β)<br />

Another interpolant operator is defined in relation to the zeroes of dxP n . Thus<br />

for n ≥ 1, let us consider Ĩw,n : C0 ([−1,1]) → Pn to be the operator associating to the<br />

function f, the unique polynomial pn ∈ Pn satisfying pn(η (n)<br />

j ) = f(η (n)<br />

j ), 0 ≤ j ≤ n.<br />

The operator is linear and we get<br />

(3.3.3) Ĩw,np = p, ∀p ∈ Pn.<br />

For Laguerre polynomials Ĩw,n : C0 (n)<br />

([0,+∞[) → Pn−1 is characterized by ( Ĩw,nf)(η j )<br />

= f(η (n)<br />

j ), 0 ≤ j ≤ n − 1, where η (n)<br />

0 = 0.<br />

As already mentioned in section 2.4, the analysis when n tends to +∞, of the<br />

operators introduced above, is one of the primary issues of chapter six.<br />

3.4 Gauss integration formulas<br />

We have reached the crucial subject of this chapter, that is the formulas for numerical<br />

integration. Given a polynomial p ∈ Pn, one is concerned w<strong>it</strong>h finding the value of<br />

<br />

I pwdx. If the Fourier expansion of p = n−1 k=0 ckuk is known, the answer has been<br />

given in chapter two, i.e., <br />

pwdx = c0<br />

I I u0wdx. On the other hand, when p is known<br />

through <strong>it</strong>s values at the zeroes of un, <strong>it</strong> is sufficient to integrate (3.2.2) obtaining<br />

(3.4.1)<br />

where w (n)<br />

j := <br />

I l(n) j<br />

<br />

I<br />

pw dx =<br />

n<br />

j=1<br />

p(ξ (n)<br />

j ) w (n)<br />

j ,<br />

wdx, 1 ≤ j ≤ n, are the weights of the integration formula.

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

Saved successfully!

Ooh no, something went wrong!