1. Not finding help here? Sign up for a free 30min 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!

Convergence of improper integrals and logs

  1. Mar 23, 2009 #1
    1. The problem statement, all variables and given/known data

    If [tex] \int_a^{\infty} \frac{f(x)}{x} dx [/tex] converges for any postiive value of a, and if f(x) tends to a limit L as x approaches 0, show that [tex] \int_0^{\infty} \frac{f(\alpha x) - f(\beta x)}{x} dx [/tex] converges for alpha and beta positive and has the value L*ln(beta/alpha).





    2. Relevant equations



    3. The attempt at a solution

    First we show it converges. Let [tex] g(x) = \frac{\alpha f(x)}{x} [/tex]. We know that the integral from a to infinity of f(x)/x converges, and so we set this equal to M and so [tex] \int_a^{\infty} g(x) dx = \alpha M [/tex]. Now, [tex]\int_a^{\infty} g(x) dx = \lim_{\epsilon \rightarrow \infty} \int_a^{\epsilon} g(x) dx = \lim_{\epsilon \rightarrow \infty} \int_{\frac{a}{\alpha}}^{\frac{\epsilon}{\alpha}} g(\alpha x)} dx = \lim_{\epsilon \rightarrow \infty} \int_{\frac{a}{\alpha}}^{\frac{\epsilon}{\alpha}} \frac{f(\alpha x)}{x} dx[/tex] Since as epsilon tends to infinity so does epsilon/beta, we have [tex] \int_{\frac{a}{\alpha}}^{\infty} \frac{f(\alpha x)}{x} dx = \alpha M [/tex]. Now we need only show that [tex] \int_0^{\frac{a}{\alpha}} \frac{f(\alpha x)}{x} dx [/tex] exists. Since f(alpha x)/x converges from a/alpha to infinity for any positive a (since alpha is also positive then so is a/alpha), and has no infinite discontinuity at x = 0, then the integral of f(alpha x)/x from 0 to a/alpha exists. Let this value be equal to N_1. Then we have [tex] \int_0^{\infty} \frac{f(\alpha x)}{x} dx = \alpha M + N_1[/tex] (alphaM + N_1). Through a similar argument, we can show that the integral from 0 to infinity of f(beta x)/x exists and has the value [tex] \beta M + N_2 [/tex] (betaM + N_2).

    I don't know how to show that [tex]\int_0^{\infty} \frac{f(\alpha x) - f(\beta x)}{x} dx = L ln(\frac{\beta}{\alpha}) [/tex] though. Any hints? Please only very small hints

    Edit: What I wrote in the brackets towards the end is what I really meant to write. The latex wouldn't let me edit it.
     
    Last edited: Mar 23, 2009
  2. jcsd
  3. Mar 23, 2009 #2
    Oh shoot, I meant to post this in the Calculus and Beyond forum. If you are a mod, can you please move it over there?
     
  4. Mar 24, 2009 #3
    Bump...
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Convergence of improper integrals and logs
Loading...