Welcome to our community

Be a part of something great, join today!

Poisson process derivation

  • Thread starter
  • Banned
  • #1

Poirot

Banned
Feb 15, 2012
250
Let customers arrive according to a poisson process with parameter st and let $X_{t}$ denote number of customers in the system by time t. Consider an interval [t,t+h] with h small.

Show that P(1 arrival)= sh + o[h], P(more than one arrival)=o[h] and P(no arrival)=1-sh+o[h].

I know P(1 arrival)=$she^{-sh}$ but how to get further?

Thanks
 
Last edited:

CaptainBlack

Well-known member
Jan 26, 2012
890
Let customers arrive according to a poisson process with parameter st and let $X_{t}$ denote number of customers in the system by time t. Consider an interval [t,t+h] with h small.

Show that P(1 arrival)= sh + o[h], P(more than one arrival)=o[h] and P(no arrival)=1-sh+o[h].

I know P(1 arrival)=$she^{-sh}$ but how to get further?

Thanks
We have a Poisson process with mean arrival rate \(s\), then the number of arrivals in an interval of length \(h\) is \(N\) and \( N \sim P(sh)\).

Then:

\( \displaystyle p(N=1)=sh\; e^{-sh} =sh \left( 1-sh + \frac{s^2h^2}{2}-... \right) = sh(1 + R(sh))=sh + shR(sh)\)

where the \(|R(sh)|\) for small positive \(h\) is bounded by \(sh\) (the truncation error for an alternating series of terms of decreasing absolute value is bounded by the absolute value of the first neglected term), so:

\( \displaystyle |sh \;R(sh)| < s^2h^2 = o ( h)\)

since \( \lim_{h \to 0} (s^2 h^2)/h = 0 \).

Now do the no arrivals case, then \(p(N>1)=1-(p(N=0)+p(N=1))\)

CB
 
  • Thread starter
  • Banned
  • #3

Poirot

Banned
Feb 15, 2012
250
Thanks, but can you explain what R(sh) is ?
 

CaptainBlack

Well-known member
Jan 26, 2012
890
Thanks, but can you explain what R(sh) is ?
It is the remainder when the series is truncated after the first term, or the sum of the tail of the infinite series (they are the same thing).

CB
 
  • Thread starter
  • Banned
  • #5

Poirot

Banned
Feb 15, 2012
250
Ok I can easily do the rest now. Thanks