0

I'm studying the following example but unable to understand a few things:

Assume all particles act independently of each other, and the probability that a particle produces $k$ offsprings is the same for all particles. Denote this probability with $f_k$, $k=0,1,2...,$ and the population size of the $t$-th generation by $X_t$, viewing $t = 0,1,...$ as time moment.

Goal is to calculate extinction probability. Let $u_t = P(X_t = 0)$ and observe that $u_t <= u_{t+1}$. So sequence ${u_t}$ is non-decreasing, all $u_t <= 1$ and hence, there is exists a number $u = \lim_{t \rightarrow \infty} u_t$ which we will call an extinction probability.

Suppose, for a while, that $X_0 = 1$ and consider the events $A$ = {the population will extinct} and $B_k$ = {the original particle has produced $k$ offsprings}, $k=1,2,...$

Since the particles act independently, once a particle has been born, the extinction probability for the part of the population generated only by this particle is the same $u$. If the original particle has produced exactly $k$ offsprings, the whole population will disappear if and only if all branches corresponding to the k original offsprings disappear. Hence, the extinction probability for the whole population is $u^k$. The formula for total probability

$u = P(A) = \sum_{k=0}^\infty P(A|B_k)P(B_k) = \sum_{k=0}^\infty u^kf_k = \sum_{k=0}^\infty f_ku^k$

Question 1:

I don't quite understand how $u$ appear on the left hand side of the expression above and on the right hand side of it.

Question 2:

I'm confused by the statement that: once a particle has been born, the extinction probability for the part of the population generated only by this particle is the same u...Hence, the extinction probability for the whole population is $u^k$. My confusion stems from if $u_t = P(X_t = 0)$ where $u_t$ is the probability of extinction of whole population at time t and $u = \lim_{t \rightarrow \infty} u_t$, then how does $u_t$ relate to $u^k$.

tkj80
  • 347
  • Do particles die off? – Joseph Eck Jun 20 '18 at 01:25
  • @JosephEck: Yes. In the next generation, only the offspring of the previous generation's particles exist; the previous generation's particles themselves are gone. – joriki Jun 20 '18 at 01:51
  • @tkj80: I don't quite understand your questions. I suspect that Question $1$ will be resolved once Question $2$ is resolved. In Question $2$, what exactly do you mean by "how does $u_t$ relate to $u^k$"? Do you understand how it relates to $u$? And do you understand why the probability of all $k$ offspring eventually going extinct is $u^k$? – joriki Jun 20 '18 at 01:54
  • 2
    Have you seen this? – Joseph Eck Jun 20 '18 at 02:01

0 Answers0