Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $X_1,\dots,X_n$ be independent exponentially distributed random variables with respective rates $\lambda_1,\dots,\lambda_n$. According to wikipedia, the following property holds $$ Pr[ X_k = \min\{X_1,\dots,X_n\} ] = \frac{\lambda_k}{\lambda_1+\dots+\lambda_n}. $$ Is there a simple proof?

share|improve this question
    
Say do it for $X_1$. If you prefer to avoid many variables, note the easy fact that $Y=\min(X_2,\dots,X_n)$ has exponential distribution parameter $\lambda_2+\cdots+\lambda_n$. So now we need the probability that $X_1\lt Y$. This is straightforward $2$ variable integration, I have written out details on MSE at least twice, as have others, but of course cannot find it. –  André Nicolas Jul 16 '13 at 3:03
add comment

1 Answer

Note that $$ \begin{align*} P(X_k=\min\{X_1,\ldots,X_n\})&=\int_0^{\infty}\left[\prod_{\substack{1\leq j\leq n\\j\neq k}}\int_{x_k}^{\infty}\lambda_je^{-\lambda_jx_j}\,dx_j\right]\lambda_k e^{-\lambda_k x_k}\,dx_k\\ &=\int_0^{\infty}\left[\prod_{\substack{1\leq j\leq n\\j\neq k}}e^{-\lambda_j x_k}\right]\lambda_ke^{-\lambda_kx_k}\,dx_k\\ &=\int_0^{\infty}\lambda_k e^{-(\lambda_1+\cdots+\lambda_n)x_k}\,dx_k\\ &=\frac{\lambda_k}{\lambda_1+\cdots+\lambda_n}. \end{align*} $$ To get to the first step, we just noticed that the event $\{X_k=\min\{X_1,\ldots,X_n\}\}$ is precisely the event that $X_j\geq X_k$ for all $j$; so, given the value of $X_k$ is $x_k$, it is the event that $X_j\geq x_k$ for all $j\neq k$. We then use independence of the variables to say that their joint density function is simply the product of the density functions, and break up over multiplication.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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