# Orthogonality Problem (From Fourier Analysis Text)

1. Apr 7, 2013

### GridironCPJ

Hello all,

I'm working through a majority of the problems in "A First Course in Wavelets with Fourier Analysis" and have stumbled upon a problem I'm having difficulty with. Please view the PDF attachment, it shows the problem and what I have done with it so far.

Once you have seen the problem and my work so far, what do you suggest I do for the next step? I thought about using the definition of the space D and setting everything less than or equal to the double integral of 1^(j+k_/2, which is just the double integral of 1 and you get xy, then I'm at a dead end, so perhaps that's not the correct route. Your ideas would be greatly appreciated.

File size:
192.7 KB
Views:
70
2. Apr 7, 2013

### AlephZero

Before you dive into the math, think what these functions look like geometrically. The region is the unit circle....

You should be able to see "why" it is true geometrically, and therefore see how to prove it before you start writing down integrals.

(Hint: the word "circle" is in bold)

3. Apr 7, 2013

### Bacle2

Another approach: You may also want to consider using the Cauchy-Goursat theorem, if you're familiar with it.

Think also of the exponential rep. e. Maybe this is what Aleph was suggesting.

And, sorry to nitpick,but you're not really letting your θn's be in L2(D); they are there already; you're letting your θn's be given by (x+iy)n

EDIT: Cauchy-Goursat does not apply here, since the product function is not analytic.

Last edited: Apr 8, 2013
4. Apr 8, 2013

### GridironCPJ

Yes, I realized earlier that my expansion was incorrect. I used polar coordinates, so we start with

∫∫[(rcosθ+irsinθ)^j(rcosθ-rsinθ)^k]rdrdθ

From here I integrated with respect to r first and then I pulled in out and was left with

[r^(j+k+2)/j+k+2]∫(cosθ+isinθ)^j(cosθ-isinθ)^k dθ

where j≠k. I am not sure exactly how to proceed. Is there an expansion of (x+iy)^j(x-iy)^k that I am not thinking of? The value in the integrand will have the form cosθ^j+k+" "+/-isinθ^j+k, where the " " represents combinations of sinθcosθ's, which integrate to zero. So I'm left with the first and last terms, which should be zero if θ=∏, correct?

If my method or logic is not optimal, please point me in the right direction.

5. Apr 10, 2013