Branching Processes

MAT 4305 - Stochastic Processes In this section we consider a class of Markov chains, known as
Lecture 3: Branching Processes branching processes, which have a wide variety of applications in the
biological, sociological, and engineering sciences.
Consider a population consisting of individuals able to produce
offspring of the same kind.
Suppose that each individual will, by the end of its lifetime, have
Department of Statistics and Computer Science
University of Peradeniya produced j new offspring with probability Pj , j ≥ 0, independently of
the numbers produced by other individuals.
We suppose that Pj < 1 for all j ≥ 0.

MAT 4305 Stochastic Processes 1 / 13 MAT 4305 Stochastic Processes 2 / 13

Definition
(The size of the zeroth generation)The number of individuals initially
present, denoted by X0 , is called the size of the zeroth generation.
Note that state 0 is a recurrent state, since clearly P00 = 1.
Also, if P0 > 0, all other states are transient.
All offspring of the zeroth generation constitute the first generation
and their number is denoted by X1 . This follows since Pi0 = P0i , which implies that starting with i
individuals there is a positive probability of at least P0i that no later
X0 =The size of the zeroth generation generation will ever consist of i individuals.
X1 =The size of the first generation Since any finite set of transient states {1, 2, . . . , n} will be visited
. only finitely often, this leads to the important conclusion that, if
. P0 > 0, then the population will either die out or its size will
. converge to infinity.
Xn =The size of the nth generation

It follows that {Xn , n = 0, 1, . . .} is a Markov chain having as its

state space the set of nonnegative integers.

MAT 4305 Stochastic Processes 3 / 13 MAT 4305 Stochastic Processes 4 / 13

Mean number of offspring of a single individual
Let us suppose that initially there is a single individual present. i. e.
∞ X0 = 1.
µ= ∑ jPj (1) We calculate E [Xn ] and Var (Xn ) by first noting that we may write
j=0
n−1
Xn = ∑ Zi
i=1

Variance of the number of offspring produced by a single individual where Zi represents the number of offspring of the i th individual of the
(n−1)th generation.
∞
σ2 = ∑ (j − µ)2 Pj (2) By conditioning on Xn−1 , we obtain E [Xn ].
j=0

MAT 4305 Stochastic Processes 5 / 13 MAT 4305 Stochastic Processes 6 / 13

Since E [X0 ] = 1, the preceding yields

The conditional variance formula now yields

Var (Xn ) = E [Var (Xn |Xn−1 )] + Var (E [Xn |Xn−1 ])

E [Xn ] = µE [Xn−1 ] = µ n

Similarly, Var (Xn ) may be obtained by using the conditional variance

formula

Var (Xn ) = E [Var (Xn |Xn−1 )] + Var (E [Xn |Xn−1 ]) (3)

Now, given Xn−1 , Xn is just the sum of Xn−1 independent random

variables each having the distribution {Pj , j ≥ 0}. Hence,
= σ 2 (µ n−1 + µ n + ... + µ 2n−2 ) since Var (X0 ) = 0
E [Xn |Xn−1 ] = Xn−1 µ, Var (Xn |Xn−1 ) = Xn−1 σ 2

MAT 4305 Stochastic Processes 7 / 13 MAT 4305 Stochastic Processes 8 / 13

 Determining the value of π0

Therefore, 1 π0 = 1 if µ < 1

Proof.
(  n

σ 2 µ n−1 1−µ
1−µ if µ 6= 1
j=0 jP(Xn = j) ≥ ∑j=0 1 . P(Xn = j) = P(Xn ≥ 1)
Var (Xn ) = µ n = E [Xn ] = ∑∞ ∞
nσ 2 if µ = 1
Since µ n → 0 when µ < 1, it follows that P {Xn ≥ 1} → 0, and hence
P {Xn = 0} → 1.

2 π0 = 1 if µ = 1
Let π0 denote the probability that the population will eventually die out
(under the assumption thatX0 = 1). More formally, 3 π0 < 1 if µ > 1 where π0 satisfies
∞
π0 = limn→∞ P {Xn = 0/X0 = 1} π0 = ∑ π0j Pj (4)
j=0

In fact when µ > 1, it can be shown that π0 is the smallest positive

number satisfying Equation (4).

MAT 4305 Stochastic Processes 9 / 13 MAT 4305 Stochastic Processes 10 / 13

Example Example
1 If P0 = 12 , P1 = 1
4 , P2 = 1
4 , then determine π0 . In previous Examples, what is the probability that the population will die
If P0 = out if it initially consists of n individuals?
4, , P2 = , then determine π0 .
1 1 1
2 P1 = 4 2

MAT 4305 Stochastic Processes 11 / 13 MAT 4305 Stochastic Processes 12 / 13

Example
Consider a branching process {Xn : n = 0, 1, 2, ...}, starting with one
particle: X0 = 1. The number of offsprings Z of one particle has distribution

0.1, for i = 0,


P(Z = i) = 0.7, for i = 1,
0.2, for i = 2.



1 Calculate the probability that the population eventually dies out and
show that for n ≥ 1,
E (Xn ) = (1.1)n ,
Var [Xn ] = (0.29) (1.1)n−1 + (1.1)n + ... + (1.1)2n−2 .
 

2 Now assume that X0 = k, for some arbitrary positive integer k > 1,

instead of X0 = 1. In this case, give again E (Xn ), Var [Xn ] and the
probability that the population eventually dies out.

MAT 4305 Stochastic Processes 13 / 13