Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Function to satisfy this integral for all parameters?

  1. Jan 23, 2008 #1
    I wonder:

    Is it possible to find a function f(x) which satisfies this integral for *arbitrary* parameters a,b (but maybe of known domain)?

    [tex]\int_{-\infty}^\infty e^{-x^2}f(ax+b)\mathrm{d}x=0[/tex]

    I want [tex]f(\pm\infty)=0[/tex] and I can deduce
    [tex]\int_{-\infty}^\infty x^n e^{-x^2}f^{(m)}(ax+b)\mathrm{d}x=0[/tex]

    It would have to be oscillary I guess?!
    Last edited: Jan 23, 2008
  2. jcsd
  3. Jan 24, 2008 #2
    Yes, and no. Obviously, f(x)=0 satisfies the criterion, but this is probably not what you were looking for. If you do a change of variable in the expression, you'll see that the integrand is exp(-((x-b)/a)^2)*f(x), i.e., a Gaussian times some function. However, if you let a and b be arbitrary, you're allowing the Gaussian's position and width to vary as well, and by making a arbitrarily large, you can make the Gaussian as thin as you want. The net effect is that the integral would "pick out" the value of f(b), so unless f is 0 everywhere, the integral cannot be 0 everywhere.

    This, of course, assumes that f is well-behaved. Any function which did satisfy the criterion would be nowhere continuous, as it would have to oscillate infinitely at every point.
  4. Jan 24, 2008 #3
    And I an estimate for b and a is known? Say it's in a range around some a0 and b0.
  5. Jan 25, 2008 #4


    User Avatar
    Homework Helper
    Gold Member

    Since coming here I have found questions I could answer, but from the other answers it looked like I hadn't understood the question. :redface: With that reservation:

    Firstly, why f(ax + b) ? Call it g(x). If you can get that you can translate into a function of (ax + b) afterwards.

    Now [tex] e^{-x^2}[/tex] is symmetrical about x = 0. So if g is any continuous and finite function of x that is antisymmetrical i.e. g(x) = - g(-x) , and defined from - to + infinity, it will satisfy your requirement won't it? Simplest example g(x) = x.

    If this shows I have not understood the question (guess no-one would ask one that easy) someone please tell me as other times I was left suspended.:confused::smile:
    Last edited: Jan 26, 2008
  6. Jan 27, 2008 #5


    User Avatar
    Homework Helper
    Gold Member


    :redface: sorry, g(x) = x doesn't satisfy your [tex]f(\pm\infty)=0[/tex] condition, but you can easily find one that does, e.g. any function symmetrical about x = 0 that does, multiplied by x.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook