Professional Documents
Culture Documents
Branching - Process
Branching - Process
20.1 INTRODUCTION
the ability to produce
Let us consider a group of similar objects which havenurnber
similar objects, either offsprings or children. The total of such cbjects
initially present are the members of the initial generation. The total number of
members or children born to the mermbers of initial generation comprises of
the first generation. Likewise the children to the first generation conprises the
members of the second generation and so on. The branching process can be
applied to compute the chance of extinction, size of a given generation and
the total size of population.
So far as the historical development of branching process in concerned,
it is believed that the process must have originated from a problem posSsed by
Francis Galton long back in 1873. The problem is of the following type :
"A large nation of whom we will only concern ourselves with adult
males,Ninnumber, and whoeach bear separate surnames, colonise a district.
Their law of population is such that, in each generation, ag percent of the
adult males have no male children who reach adult life; a, have one such
male children who reach adult life; a, have two; and so on upto az who have
five. To find () what proportion of the surnames will have become extinct
after r generations and (i) how many instances there will be of the san
surname being held by n persons."
The first attempt to solve the problen was by Watson. However, Hevde
and Seneta later pointed out that a similar problem was also posed by
Bienayme' in 1845, atleast 28 years earlier to Galton.
Whoever has initiated the problem of branching process, the tool is
extremely useful for modelling severalphysical process such as chemical and
nuclear chain reaction. This also serves as a crude model for population
growth.
Let N(n, i) be the number of childrern born to the h individual of the nth
generation. The basic assumption ia that the random variables N(n, i), (n e N,
i=1,2, ..) are independent and identically distributed.
Let, P:= P[N (n, i) =k], k=0,1, ...
We define the probability generation function of the
same as,
G(s) = , sp, k=0, 1,2, .. and 0s1.
k=0
Let ,be a
If X, = 0, the random variable representing the size of the n" generation.
accordingly,X,1population
is said to become extinct in the nth
=0, Xy= 0, . generation,
But, if X,, =
i0then,
Thus,we have Xn+1 = N(1, 1) +N(n, 2) +...+ N(n, i)
Xyl =0 if X, = 0
= N(n, 1) +N(n, 2) + ... + N(n, X,) if X,>0
Thus, (Xprocess.
a branching n =0, 1, 2, ...] is a Markov
chain.
Such a stochastic prOcess is
Suppose that the size of the initial generation is 1,
So,we can i.e., Xo = 1.
write,
i=1
where Z, = number of off springs of the ith individual
Let us assume that Z,s are iid random variables with distribution {p
Now, the probability generating function of the branching process is
given by,
P,(s) = P(X,=k<s
k=0
( öP
, (6))
Po
Po 45°
45°
Fig
q, <E<1.
P(8))(0 =8and
y=sininduction,
(8, <P8) å
pointabove in
E
by
=P{0) value
and so
another when,
s1
(1) and a
1) 4, 8<1. is
at (0, Then,
6, there unity
1-s P =sin = P(6) = than
y s= <s<ô. qthat, 1) P(1)-P()
s 1. y <
intersects P(q) (8, 1-8
P(s) ns line0 interval less
so P()>1.
1- = thein
s - S, so, root1.P(1)>
I)-P)= (1) below > q,m+1 = P(E)=monotone
r)-P)P' P(s) Pls) . the the
1-$ when =v lies
get,
à<s<land It in is
curve we consider q
=l convex, is that
<Ó, derivative
Lin1 . The find
Lim- 0 being in 4,, us
: let we
II <s that Hence,
Ihen Thus, Case P(s) Now, the Thus,
that ASsuming such
that,
suh i.e., as
S), and