3
$\begingroup$

I think I'm confused by the fact that they have similar names and both involve i.i.d. exponential RVs. From what I understand,

Poisson distribution: the probability of a particular number of events in a fixed interval. And the inter-arrival times are i.i.d. exponential RVs?

Poisson process: is a sequence of events such that their arrival times are i.i.d. exponential RVs.

So, what's the difference between these? Is there any relation between the two? If someone could give me a bit of a dumbed down intuitive understanding that would probably go a lot farther for me than a very technical treatment. Any help is appreciated, thanks.

$\endgroup$

2 Answers 2

4
$\begingroup$

There most certainly is a relation between the two!

In general, a stochastic process is simply an indexed collection of random variables defined on a probability space $(\Omega, \mathcal F,\mathbb P)$, e.g. $N=\{N(t):t\in I\}$ for some index set $I$. A Poisson process is one of the most well-known examples of an arrival process, which satisfies the following properties (with probability $1$, or "almost surely"):

  1. $N(0) = 0$
  2. If $s<t$, then $N(s)\leqslant N(t)$.
  3. The map $t\mapsto N(t)$ is right-continuous.
  4. If $\lim_{t\to t_0^-} N(t)<N(t_0)$, then $N(t_0)=N(t)+1$.

The fourth condition describes what are called jumps, that is, the random times at which arrivals occur. If we set $T_0=0$ and $$T_{n+1} = \inf\{t>T_n : N(t)>N(T_n)\} $$ for $n=0,1,2,\ldots$ then $T_n$ are the arrival times of the process $N(t)$. The times between arrivals $T_{n+1}-T_n$ are called interarrival times. A Poisson process is an arrival process which also satisfies the following conditions:

  1. For any $s,t\geqslant 0$ and $0\leqslant u_1<\cdots<u_n\leqslant t$, $N(t+s)-N(t)$ is independent of $\{N(u_1), N(u_2),\ldots, N(u_n)\}$.
  2. For any $s,t\geqslant 0$, $N(s+t)-N(t)$ is independent of $t$.

These are referred to as independent increments and stationary increments, respectively. From these conditions (and assuming that $N(t)$ is not identically zero) it can be shown that there exists $\lambda>0$ such that $\mathbb P(N(t)=k) = \frac{e^{-\lambda t}(\lambda t)^k}{k!}$ for all $t>0$, $k=0,1,2,\ldots$ (for brevity, I will omit the argument, which can be found in any text on stochastic processes). This $\lambda$ is the rate or intensity of the Poisson process.

From there, we can see that for $s,t> 0$ and $k=0,1,2,\ldots$, $$\mathbb P(N(s+t)-N(t)=k) = \frac{e^{-\lambda s}(\lambda s)^k}{k!}, $$ that is, the distribution of arrivals in a time interval of length $s$ follows a Poisson distribution with rate $\lambda s$. To derive the interarrival distribution, note that for for any $n\geqslant 0$ and $t>0$, $$\mathbb P(T_{n+1}-T_n>t) = \mathbb P(T_1>t) $$ due to stationary increments, and $$\mathbb P(T_1>t) = \mathbb P(N(t) = 0) = e^{-\lambda t}, $$ hence the time between arrivals is exponentially distributed with rate $\lambda$.

$\endgroup$
3
$\begingroup$

A random process is, unlike a probability distribution $[1]$, a family of random variables $\{X(t),\: t \in T\}$ defined on a given probability space, indexed by the parameter $t$, where $t$ varies over an index set $T$.

A random process is said to be a counting process if $X(t)$ represents the total number of events that have occurred in the interval $(0,t)$.

A counting process is said to be a Poisson process with rate $\lambda$ if

  1. $X(0) = 0$.

  2. $X(t)$ has independent increments.

  3. The number of events in any interval of length $t$ is Poisson distributed $[2]$ with mean $\lambda t$; that is, for all $s, t > 0$,

$$P[X(t+s)-X(s) = n] = e^{-\lambda t}\frac{(\lambda t)^n}{n!},\:\:\: n = 0,1,2...$$


"So, what's the difference between these?" $[1]$

"Is there any relation between the two?" $[2]$

$\endgroup$

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.