Assume that calls arrive at a call centre according to a Poisson arrival process with a rate of 15 calls per hour. For $0 ≤ s < t$, let $N(s,t)$ denote the number of calls which arrive between time $s$ and $t$ where time is measured in hours.

What is $E(N(1,4)N(2,6))$?

I am confused by this because I know $E(XY) = E(X) E(Y)$ if X and Y are independent but in this case the periods of time overlap and are not independent.


You can split the intervals up:

$$N(1,4)=N(1,2)+N(2,4)\qquad\text{ and }\qquad N(2,6)=N(2,4)+N(4,6).$$


\begin{eqnarray*} E(N(1,4)N(2,6)) &=& E([N(1,2)+N(2,4)]\;N(2,6))) \\ &=& E(N(1,2)N(2,6) + N(2,4)\;[N(2,4)+N(4,6)]) \\ &=& E(N(1,2)N(2,6) + N(2,4)^2 + N(2,4)N(4,6)) \\ &=& E(N(1,2))E(N(2,6)) + E(N(2,4)^2) + E(N(2,4))E(N(4,6)) \\ &=& \lambda\cdot 4\lambda + (2\lambda)^2 + 2\lambda + 2\lambda\cdot 2\lambda \\ &&\qquad\qquad\qquad\qquad\text{since $E(N(s,t)^2)=((t-s)\lambda)^2+(t-s)\lambda$} \\ &=& 12\lambda^2 + 2\lambda \\ &=& 2730 \quad\text{with $\lambda=15$}. \end{eqnarray*}

  • $\begingroup$ thank you! I have not come across this before. $\endgroup$ – chronologos Mar 18 '15 at 14:53

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.