Chapter 8. Chebyshev spectral methods
Chapter 8. Chebyshev spectral methods
Chapter 8. Chebyshev spectral methods
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>8.</strong>3. CHEBYSHEV DIFFERENTIATION BY THE FFT TREFETHEN 1994 275<br />
ALGORITHM FOR CHEBYSHEV DIFFERENTIATION<br />
1. Given data fv j g de ned at the <strong>Chebyshev</strong> points fx j g, 0 j N, think of the same<br />
data as being de ned at the equally spaced points f j g in [0 ].<br />
2. (FFT) Find the coe cients fa ng of the trigonometric polynomial<br />
that interpolates fv j g at f j g.<br />
3. (FFT) Compute the derivative<br />
dq<br />
d<br />
NX<br />
q( )= an cos n (8:3:12)<br />
n=0<br />
NX<br />
= ; nan sin n : (8:3:13)<br />
n=0<br />
4. Change variables to obtain the derivative with respect to x:<br />
dq dq d<br />
=<br />
dx d dx =<br />
NX<br />
n=0<br />
na n sin n<br />
sin<br />
NX<br />
=<br />
n=0<br />
na n sin n<br />
p 1;x 2<br />
At x = 1, i.e. =0 , L'Hopital's rule gives the special values<br />
NX<br />
: (8:3:14)<br />
dq<br />
( 1) = ( 1)<br />
dx<br />
n=0<br />
n n 2 an (8:3:15)<br />
5. Evaluate the result at the <strong>Chebyshev</strong> points:<br />
wj = dq<br />
dx (xj ): (8:3:16)<br />
Note that by (<strong>8.</strong>3.3), equation (<strong>8.</strong>3.12) can be interpreted as a linear combination of<br />
<strong>Chebyshev</strong> polynomials, and by (<strong>8.</strong>3.6), equation (<strong>8.</strong>3.14) is the corresponding linear combination<br />
of derivatives.* But of course the algorithmic content of the description above<br />
relates to the variable, for in Steps 2 and 3, we have performed Fourier <strong>spectral</strong> di erentiation<br />
exactly as in x7.3: discrete Fourier transform, multiply by i ,inverse discrete Fourier<br />
transform. Only the use of sines and cosines rather than complex exponentials, and of n<br />
instead of , has disguised the process somewhat.<br />
*orof<strong>Chebyshev</strong> polynomials Un(x) ofthe second kind.