22.06.2013 Views

Wavelet Galerkin Solutions of Ordinary Differential Equations

Wavelet Galerkin Solutions of Ordinary Differential Equations

Wavelet Galerkin Solutions of Ordinary Differential Equations

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

<strong>Wavelet</strong> <strong>Galerkin</strong> solutions 415<br />

simultaneous equations in wavelet space and not in physical space. The solution in<br />

wavelet space is then transformed back to physical space by FTT.<br />

Let the wavelet expansion u(x) at scale j be<br />

j / 2 j<br />

u(<br />

x)<br />

c 2 ( 2 x k)<br />

, k Z,<br />

(4.2)<br />

k<br />

k<br />

where ck s are periodic wavelet coefficients <strong>of</strong> u, periodic in x.<br />

j<br />

Put y 2 x so that<br />

U ( y)<br />

u(<br />

x)<br />

C ( y k),<br />

C 2<br />

k<br />

k<br />

k<br />

If d is the period <strong>of</strong> u, then U (y)<br />

and so also Ck is periodic in y with period 2 d.<br />

j<br />

j / 2<br />

j<br />

Let us discretize U(y ) at all dyadic points x y y Z<br />

<br />

2 ,<br />

U i C k<br />

ik<br />

C<br />

ik<br />

k , i 0,<br />

1,<br />

2,....,<br />

n 1.<br />

k<br />

k<br />

The matrix representation is U k<br />

* C , where k is the convolution kernal, i.e. the<br />

first column <strong>of</strong> the scaling function matrix.<br />

Similarly the wavelet expansion for f (x),<br />

j / 2 j<br />

f ( x)<br />

d 2 ( 2 x k)<br />

, k Z.<br />

(4.3)<br />

k<br />

k<br />

j / 2<br />

f ( x)<br />

Dk<br />

( y k),<br />

Dk<br />

2 dk<br />

.<br />

k<br />

F(<br />

y)<br />

<br />

And the matrix representation is<br />

F k<br />

* D.<br />

Substitute the expansions <strong>of</strong> u(x) and f (x)<br />

in (4.1) and then take inner product on<br />

both sides with ( y j),<br />

j Z .<br />

( n)<br />

Use j k<br />

<br />

( y k)<br />

(<br />

y j)<br />

dy and jk (<br />

y k)<br />

(<br />

y j)<br />

dx , we obtain<br />

k . C g.<br />

Now take Fourier Transforms<br />

Uˆ kˆ<br />

. Cˆ<br />

.<br />

Fˆ kˆ<br />

. Dˆ<br />

.<br />

kˆ . Cˆ<br />

gˆ<br />

.<br />

Subsequently, Uˆ<br />

Fˆ<br />

/ kˆ<br />

. Inverse FT will give U.<br />

Wherein . and / denote component by component multiplication and division<br />

respectively.<br />

c<br />

k<br />

.

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

Saved successfully!

Ooh no, something went wrong!