There is a random variable $r = \dfrac 1 {I(y_1<c) + I(y_2<c)}$. Both $y_1$ and $y_2$ are i.i.d. random variables with exponential distribution, so their joint distribution is $f(y_1, y_2) = \prod_{i=1}^2 \theta^{-1} e^{- \dfrac {y_i} \theta}$.

My goal is to calculate expected value $E(r)$. I tried it myself and got stuck at some point. Below $I(y<c)$ is the indicator function. This is what I have attempted so far :

$ \begin{align} E(r) &= \int_0^{\infty} \int_0^{\infty} \dfrac 1 {I(y_1<c) + I(y_2<c)} \theta^{-1} e^{- \dfrac {y_1} \theta} \theta^{-1} e^{- \dfrac {y_2} \theta} dy_1 dy_2 \\ &= \int_0^{\infty} \bigg( \int_0^{c} \dfrac 1 {1 + I(y_2<c)} \theta^{-1} e^{- \dfrac {y_1} \theta} dy_1 + \int_c^{\infty} \dfrac 1 {I(y_2<c)} \theta^{-1} e^{- \dfrac {y_1} \theta} dy_1 \bigg) \ \theta^{-1} e^{- \dfrac {y_2} \theta} dy_2 \\ &= F_{y_1}(c) \int_0^{\infty} \dfrac 1 {1 + I(y_2<c)} \theta^{-1} e^{- \dfrac {y_2} \theta} dy_2 \ + (1-F_{y_1}(c)) \int_0^{\infty} \dfrac 1 {I(y_2<c)} \theta^{-1} e^{- \dfrac {y_2} \theta} dy_2\\ &= F_{y_1}(c) \bigg(\dfrac 1 2 F_{y_2}(c) + (1-F_{y_2}(c)) \bigg) \ + (1-F_{y_1}(c) ) \bigg(F_{y_2}(c) + \int_c^{\infty} \dfrac 1 {I(y_2<c)} \theta^{-1} e^{- \dfrac {y_2} \theta} dy_2 \bigg) \end{align} $

I am stuck because $I(y_2<c) = 0$ in that integral. How should I solve this problem?


This random variable has a positive probability to be infinite, therefore its expectation is $+\infty$.

| cite | improve this answer | |

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.