Does memoryless property imply identical distribution of inter-arrival CDF's in a poisson process?

38 Views Asked by At

This is my result from finishing a problem in a stochastic process workbook:

"Since a Poisson process has the property of stationary increments, this implies that the inter-arrival times $Z_n=T_n-T_{n-1}$ are identically distributed without regard to the starting value of n. That is:

$$\Pr(Z_{n-1} \le z) = \Pr(Z_n \le z)$$

$$F_{Z_{n-1}}(z) = F_{Z_{n}}(z)$$

For inter-arrival times of a Poisson Process, the CDF of step n-1 equals CDF for step n


How does this property relate to the memoryless property for exponential random variable, namely:

  • A RV X is memoryless if X is positive, and for all real $t>0$ and $s >0$, then:

    $P[X > t +s] = P[X>t]~P[X>s]$

  • A RV X is memoryless if for any s and t, we have:

    $P(X > t+s~|~X > t) = P(X>s)$

I was just curious... because CDF and memoryless properties seem similar to each other ... but the sign is backwards...