2
$\begingroup$

I try to solve an integral equation as follow: $$ \int_{0}^{x}f(t)f(x-t)dt=e^{2x}-1, \textrm{where }f(x) \textrm{ is continuous and nonnegative.} $$ I want to solve such $f(x)$,but after trying to take the derivative of both sides of the equation, I still have no idea. The only thing I get is $f(0)=\sqrt{2}$. Anyone can help me with this?

$\endgroup$
  • $\begingroup$ Actually if we only assume $f$ is continuous, then we can't even take the derivative. So I think if we need to assume $f$ is smooth or something like this? $\endgroup$ – Bob brant Sep 21 '17 at 4:55
  • $\begingroup$ This equation has a convolutional form $f*f(x)=e^{2x}-1$ under the following restriction: $x>0$. Usually, this kind of equation is solved using Laplace transform, but here I don't see how it could give an explicit solution... $\endgroup$ – Jean Marie Sep 21 '17 at 5:06
6
$\begingroup$

With Laplace transform: $$\mathcal{L}\int_{0}^{x}f(t)f(x-t)dt=\mathcal{L}(e^{2x}-1)$$ $$(\mathcal{L}f)^2=\dfrac{1}{s-2}-\dfrac1s$$ $$\mathcal{L}(f)=\dfrac{\sqrt{2}}{\sqrt{s(s-2)}}$$ $$f(x)=\sqrt{2}e^tI_0(x)$$ where $I_0(t)=J_0(it)$ is Bessel function.

$\endgroup$
  • $\begingroup$ [+1] I had not been able to identify the last Laplace transform ! $\endgroup$ – Jean Marie Sep 21 '17 at 5:29
  • $\begingroup$ @JeanMarie Thanks. $\endgroup$ – Nosrati Sep 21 '17 at 10:17
  • $\begingroup$ Thanks a lot. I think I need to know more about the Laplace transform. $\endgroup$ – Bob brant Sep 21 '17 at 11:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.