1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Homework Help: Integral Product of Cosines to Show Orthogonality

  1. Oct 7, 2012 #1
    1. The problem statement, all variables and given/known data
    Show that (forgive me for not knowing how to use latex)

    from x=0 to x=1 of:

    ∫cos([(2n+1)(pi)/2]x)*cos([(2m+1)(pi)/2)]x) dx = 0, for m ≠ n

    2. Relevant equations
    The question tells me to use integral tables.

    3. The attempt at a solution
    Using integral tables, I got

    ∫(1/2) [cos((A-B)x) + cos((A+B)x)] dx from 0 to 1.

    A, B are equal to (2n+1)pi/2 and (2m+1)pi/2 respectively.

    However, when I evaluate the integral and I get sines, the x merely becomes a 1 and I get a little confused after that.

    Any help would be greatly appreciated.
    Last edited: Oct 7, 2012
  2. jcsd
  3. Oct 7, 2012 #2
    What is ## \int_0^1 \cos \alpha x dx ##?
  4. Oct 7, 2012 #3
    I believe it is sin(ax)/a evaluated from 0 to 1, so that would be sin(a)/a, correct?
  5. Oct 7, 2012 #4
    Exactly correct. Now substitute A + B and A - B.
  6. Oct 7, 2012 #5
    For the integration from 0 to 1, for (a+b) and (a-b) respectively I get:



    Since it is basically a sum of integrals, I can just sum these (with the accompanying 1/2), correct? But I don't know what else after that: this is where I got confused.
  7. Oct 7, 2012 #6
    There is one problem with you derivation. When A = B, you cannot integrate cos (A - B)x as sin (A - B)x / (A - B), because A - B = 0. However, cos (A - B)x = cos 0 = 1, and that can be integrated. What is the result? Note this is for m = n.

    In all the other cases, what are sin (A + B) and sin (A - B)?
  8. Oct 7, 2012 #7
    Right, okay, so if m = n, then the cosine term immediately becomes cos(0) = 1, and the integral of 1 in this case would just be the variable x. But why are we considering m = n?

    As for the second question... I guess that A+B would be a positive number and A-B would be a negative number??? In all other cases... I'm a little confused by what you mean. Hopefully I am interpreting correctly.
  9. Oct 7, 2012 #8
    We are considering m = n because A - B = 0 in the integrand is not a cosine but a constant.

    For other cases, look at what A - B and A + B really are. What is the sine of them?
  10. Oct 7, 2012 #9
    Ah! You mean the literal sine of the quantities A + B, A - B, right?

    So the sum of A and B are sums of (σ∏)/2, where σ is some constant made by the sum of two odd numbers. And the sum of any odd numbers is an even number. Which puts the A + B into the form of just σ∏. Sine of a multiple of ∏=0.

    The difference of A and B would give the same type of result, right?

    I hope I'm following correctly. I still don't know the grounds for considering m = n when the question asks to prove the case for m ≠ n. I'm guessing this will all be revealed through this work.
  11. Oct 7, 2012 #10
    I thought I got the proof, but I did not...
    Last edited: Oct 7, 2012
  12. Oct 7, 2012 #11
    Oh what the hell I'll just say it.

    Since any sum of A and B (negative sum or otherwise) gives an even number multiple of ∏, then the sin (A+B) and sin (A-B) terms would give zero.
  13. Oct 7, 2012 #12
    Exactly. Unlike the m ≠ n case.
  14. Oct 8, 2012 #13
    In the m ≠ n case though, it is just a sum of ∏/2 terms. And any sum (negative or otherwise) of ∏/2 terms gives a term of ∏, which for sin = 0. So if n = 1000 or m = 12312412, the sine of A + B or A - B would still be zero, no?
  15. Oct 8, 2012 #14
    By turning the product of the cosines into a sum of cosines, then it becomes clear that the integral in question would be zero after evaluation, I believe.
  16. Oct 8, 2012 #15
    You have proven this, but since you sound a bit unsure, I will repeat that for you :)

    Both A and B are an odd number of pi/2, so both their sum and their difference is an even number of pi/2, that is to say, some integer number of pi, the sine of which is zero. Hence the integral is zero unless m = n, when there is a constant term, which does not integrate to a sine.
  17. Oct 8, 2012 #16
    Awesome Voko, thanks so much. You did an excellent job of leading me to my own conclusion. Thanks again!

    It would be nice if I could give you a +1 or something along those lines. Great help!
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook