I recently had to solve this same problem, except $X$ and $Y$ were uniform on $[0,1]$. The joint probability distribution was uniform, so I just needed to find the proportion of the area inside the region of the unit square where $|X-Y|\leq z$ for a given $z$.

I was wondering: what should I do if $X$ and $Y$ are independent exponential random variables with parameter $\lambda$? I believe I would have to integrate $$\int_{y_0}^{y_1} \int_{x_0}^{x_1} \lambda^2 e^{-\lambda(x+y)} \, \mathrm dx \, \mathrm dy$$

but I'm not sure what the bounds would be. How should one approach this?


You can use: $$ P(\lvert X - Y \rvert < z ) = P(X - Y < z, X - Y > -z) = P(X - Y < z) - P(X - Y < z, X - Y < -z) = P(X- Y < z) - P(X-Y<-z)$$

While $X-Y$ has pdf that's quite easy to calculate - see another answer. Moreover, it's form allows us for closed-form expression for cdf. Said PDF is:

$$ f(x) = \frac{\lambda \mu}{\lambda+\mu} \cases{e^{-\mu x} & if $x > 0$\cr e^{\lambda x} & if $x < 0$\cr}$$

where $\mu, \lambda$ are parameters of $X, Y$ respectively.


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.