I consider a poisson process with rate $$\lambda$$ . Since the Inter arrival time follows exponential distribution, the expectation value of the inter arrival time, say $$E[X]= \frac{1}{\lambda}$$. If I look particularly for a time interval [0-15s]. Then,

  1. What would be the expectation value of the inter arrival time e.g., first, second, third arrivals,......? Would it be $$E[X| t<15]$$

  2. What is the expectation value of the inter arrival time if I can it is conditioned over [5s-15s] i.e., $$E[X| 5<t<15]$$?

  • $\begingroup$ To be clear: Are you looking for $\mathsf E(X\mid X\in[s;t))$ ? That is the (conditional) expected inter-arrival time when given that it occurs within a specified interval $s\leqslant X<t$. $\endgroup$ – Graham Kemp Jul 21 '17 at 9:16
  • $\begingroup$ By the way, the time until the $k$-th arrival where $k>1$ is not distributed the same as the first arrival. $\endgroup$ – Nap D. Lover Jul 21 '17 at 11:48
  • $\begingroup$ Dear @LoveTooNap29 your comment is very right. Would you please elaborate then the right distribution for the $k^{th}$ arrival where $k>1$? Thanks in advance $\endgroup$ – Hallian1990 Dec 22 '17 at 19:41
  • $\begingroup$ @Hallian1990 for $k=1,2,...$ the time until the $k$-th arrival of a Poisson process with mean parameter $\lambda$ is Erlang distributed with parameters $k$ and $\lambda$. The mean time for the $k$-th arrival is $k/\lambda$. Note, an Erlang RV is just a gamma RV with an integer shape parameter, also note when $k=1$, the Erlang distribution reduces to the exponential distribution. If you are unfamiliar with this distribution, a quick google or wiki search will remedy that. $\endgroup$ – Nap D. Lover Dec 22 '17 at 20:03

Are you looking for the expected inter arrival time when given that that the next event occurs within a particular interval?

$$\begin{align*}\mathsf E(X\mid s\leqslant X\lt t) ~&=~ \dfrac{\int_s^t \lambda x \mathrm e^{-\lambda x}\operatorname d x}{\int_s^t \lambda \mathrm e^{-\lambda x}\operatorname d x} \\[2ex] &=~ \dfrac{(\lambda s+1) \mathrm e^{-\lambda s}-(\lambda t+1)\mathrm e^{-\lambda t}}{\lambda (\mathrm e^{-\lambda s}-\mathrm e^{-\lambda t})}\end{align*}$$

  • $\begingroup$ Hi Kemp. Yes, I am looking to find the conditional expectation value of inter arrival time. I understood your point. Many thanks. I have just a short question to make it clear. for a Poisson process with an time interval e.g., [0-t] and rate lambda, if the first arrival has an expectation value of E[X1]=E[X | 0<=X<t], then for the second arrival can I say the expectation of the second is E[2]= E [X | E[X1]<=X<t ], means that the second will have expected value between E[X1] and t? $\endgroup$ – Hallian1990 Jul 21 '17 at 11:46
  • $\begingroup$ Dear Kemp, I want to calculate the expected value of the inter arrival time of the FIRST and SECOND arrivals, by considering a poisson process with rate $\lambda$ and for time $t= [0-10s]$ How to express this conditional expectation value. In other words, how much time (expected) will be elapsed until we get 2 arrivals, conditioning time this time is less than $10s$ $\endgroup$ – Hallian1990 Jul 22 '17 at 0:38

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.