3
$\begingroup$

I have two Poisson distributions with parameters $\lambda_1$ and $\lambda_2$ and two independent variables $A$ and $B$ from these two distributions. I know that

$$\begin{equation*} \mathsf P(\min(A,B)>K)=\mathsf P(A>K) \ast \mathsf P(B>K) \end{equation*}$$

I'm just wondering what $\mathsf P(\max(A,B)>K)$ is. Thanks.

$\endgroup$
4
$\begingroup$

If the maximum of two numbers is greater than some constant, then either one or the other of the numbers is greater than it;   they are not both less than or equal to it.

If the random variables are independently Poisson distributed then:

$$\begin{align} \mathsf P(\max(A,B)>K) & =\mathsf P(A>K \cup B>K) \\[1ex] & = 1 - \mathsf P(A\leq K)\mathsf P(B\leq K) \\[1ex] & = 1 - \sum_{j=0}^K\frac{\lambda_1^j e^{-\lambda_1}}{j!}\;\sum_{j=0}^K\frac{\lambda_2^j e^{-\lambda_2}}{j!} \\[1ex] & = 1 - \frac{\Gamma(K+1,\lambda_1)\Gamma(K+1,\lambda_2)}{K!^2} \end{align}$$


$\Gamma(\cdot,\cdot)$ is the Incomplete Gamma Function

$\endgroup$
  • $\begingroup$ Thank you. Is this approach also true for more than two variables? $\endgroup$ – Taban May 11 '15 at 10:53
  • $\begingroup$ Both of the Answers are applicable to several independent random variables. This one for several independent Poisson random variables. $\endgroup$ – BruceET May 12 '15 at 3:13
5
$\begingroup$

Let $W = \max(A, B)$, then $P(W > k) = 1 - P(W \le k) = 1 - P(A \le k)P(B \le k).$

$\endgroup$
  • 1
    $\begingroup$ I think $W$ became $X$ and $K$ became $k$. +1 anyway. $\endgroup$ – Nicolas May 11 '15 at 9:08
  • 1
    $\begingroup$ Right. Phone rang while proofing. $\endgroup$ – BruceET May 11 '15 at 9:09

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.