unique solution to integral functional equation

45 Views Asked by At

I have the following equation:

$f(x,y)=Pr(v\xi-y>max\{v-x,0\})$ $ $ $ $ $ $ $ $ $ $ $ $ $\forall x,y$ $ $ $ $ (1)

where $f(x,y)$ is a known function and $f(x,y) \in (0,1)$ , $v$ and $\xi$ are independent random variables with $v\sim F_v$ ($F_v$ is the CDF of $v$) and $\xi \sim F_\xi$ ($F_\xi$ is the CDF of $\xi$ with domain $[1,+\infty$) ). In the above equation (1) $F_v$ is known, whereas $F_\xi$ is unknown.

I want to understand whether there is a unique CDF $F_\xi$ solving equation (1)/ for which equation (1) holds.

How should my proof proceed? Is anyone familiar with similar cases?

1

There are 1 best solutions below

3
On BEST ANSWER

How should my proof proceed? Is anyone familiar with similar cases?

When working with probability problems, if you aren't sure how to proceed, a good question to ask is: what if my random variable is…

  • …a normal distribution (this might be tricky, because the normal CDF has no elementary closed form),
  • …a Cauchy distribution, or
  • …a constant (a.s.)?

In your case, suppose $v=0$ a.s. Then $$f(x,y)=\mathbb{P}[{0\cdot\xi-y>\max(v-x,0)}]=\mathbb{P}[-y>\max(v-x,0)]$$ which has no dependence on $\xi$. Consequently it cannot determine $\xi$ uniquely.