3
$\begingroup$

I've got the following problem I'm working on which is related to some of my research:

Solve:

$f(x) = \int_{-\infty}^x G(x,y)f(y)f(x-y)dy$

for f, given $G$ which has whatever smoothness properties you might need.

There are two possible simplifications that can be made regarding G: either

1: $G(x,y) \equiv 0$ when $y>x$ and thus we continue the integral up to $\infty$

or

2: $G$ is symmetric, $G(x,y) = G(y,x)$ (but the integral must extend to x, not $\infty$)

These are the only bounds I have so far to put on G.

Hoping you guys might have some ideas. This question comes about in a few problems related to the study of the structure of solids.

-edit

I'm not really looking for a general solution, since its highly unlikely it will be possible. I'm really interested in finding particular solutions, assumptions, or techniques that might be helpful in understanding the problem.

--

$\endgroup$
4
  • 1
    $\begingroup$ I'm confused: it sounds like you are saying $G(x,y) = G(y,x)$ and $G(x,y) = 0$ for $y>x$. That would mean $G(x,y)=0$ unless $x=y$ and therefore $F(x) = 0$ for all $x$. $\endgroup$
    – Noah Stein
    May 29, 2012 at 21:00
  • $\begingroup$ Written in this way is very difficult to help. What is $G$ and what is unknown, $f$ or $F$? $\endgroup$
    – Jon
    May 29, 2012 at 21:05
  • 1
    $\begingroup$ I can't imagine a reason why there could be non-zero solutions... The map $f\mapsto f - Q(f)$ with a quadratic Q is locally invertible at 0, so that the solution f=0 should be isolated. On the other hand, it is difficult to imagine an invariant set which is not a nbd of 0. $\endgroup$ May 30, 2012 at 5:14
  • $\begingroup$ Pietro, there do exist non-zero solutions, for example the Weibull distribution. If f(x) = 0 w/ x<0, $G(x,y) = C*x^{\gamma}$ we can pull it out. $\endgroup$
    – Chris
    May 30, 2012 at 17:01

1 Answer 1

3
$\begingroup$

The simplest approach one can think is an iterative one. Let us consider the given equation $$ f(x)=\int_{-\infty}^xG(x,y)f(y)f(x-y)dy. $$ Now, we assume as a first iterate $f^{(0)}(x)=1$ and so $$ f^{(1)}(x)=\int_{-\infty}^xdyG(x,y) $$

$$ f^{(2)}(x)=\int_{-\infty}^xdyG(x,y)\int_{-\infty}^ydwG(y,w)\int_{-\infty}^{x-y}dzG(x-y,z) $$

and so on. One is granted the existence of the n-th iterate provided the integral of $G$ exists and is properly bounded.

$\endgroup$
2
  • $\begingroup$ We also need the integrand to be Lipschitz, right? $\endgroup$
    – Chris
    May 30, 2012 at 17:10
  • $\begingroup$ Chris, I think so. $\endgroup$
    – Jon
    May 30, 2012 at 19:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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