1
$\begingroup$

We know that $X_1,X_2$ and $X_3$ are three independent exponential random variables, with rates $\lambda_1,\lambda_2$, $\lambda_3$ respectively. We want to calculate $$E[\max{X_i}| X_1<X_2<X_3]$$ The answer is equal to $$\frac{1}{\lambda_3}+\frac{1}{\lambda_2+\lambda_3}+\frac{1}{\lambda_1+\lambda_2+\lambda_3}$$ I have read the solution for this problem, and it seems to be correct for me, but I cannot understand why we cannot explain the following solution for this problem:

As we know that $X_1<X_2<X_3$, so $\max{X_i}=X_3$, hence we have $E[\max{X_i}| X_1<X_2<X_3]=E[X_3]=\frac{1}{\lambda_3}$

Any help?

$\endgroup$
2
$\begingroup$

Following your logic rigorously, one gets the correct identity $$E[\max{X_i}\mid X_1<X_2<X_3]=E[X_3\mid X_1<X_2<X_3].$$ Note that the RHS is definitely not $E[X_3]$.

$\endgroup$
1
$\begingroup$

Intuitively speaking, the solution you suggest simply ignores the constraint $X_1<X_2<X_3$. Yes, you calculate the expected value of $X_3$, but you include in your calculation events where $X_3$ is not the maximum.

$\endgroup$
  • $\begingroup$ I cannot understand the last part of your answer "but you include in your calculation events where $X_3$ is not the maximum". Can you explain more about your answer? $\endgroup$ – CLAUDE Nov 30 '14 at 14:23

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.