STAT 333 - Problem Set 7
Exercises marked with an asterisk ∗ are considered to be recommended exercises. The non-
asterisk exercises are intended to serve as extra practice exercises.
Section 3.4
Exercise 3.4.1. Consider a Galton-Watson Branching Process with X0 = 1 and µ > 1.
Use mathematical induction to showPthat z ∗ ≥ P (Xn = 0), n ∈ N, where z = z ∗ is any
non-negative solution satisfying z = ∞ j
j=0 z αj .
Exercise 3.4.2.∗ For a Galton-Watson Branching Process having X0 = 1, calculate the prob-
ability that the population will eventually die out when the number of offspring (produced
by any one individual) has probability distribution {αm }∞
m=0 given by
(a) α0 = 1/2, α1 = 1/5, α2 = 1/5, α3 = 1/10.
(b) α0 = 2/5, α1 = 3/10, α2 = 1/5, α3 = 1/10.
(c) α0 = 1/4, α1 = 1/3, α2 = 1/4, α3 = 1/6.
(d) α0 = 1/6, α1 = 1/3, α2 = 5/12, α3 = 1/12.
(e) α0 = 3/7, α1 = 1/35, α2 = 12/35, α3 = 2/35, α4 = 1/7.
Exercise 3.4.3.∗ Consider a Galton-Watson Branching Process {Xn , n ∈ N} with X0 = 1.
Suppose that the number of offspring (produced by any one individual) has probability
distribution {αm }∞
m=0 which satisfies
α0 = 1 − q, α1 = (1 − q)q, α2 = q 2 , αm = 0 ∀ m ≥ 3 where 0 < q < 1.
(a) Find a condition on q which ensures the probability that the population will eventually
die out is strictly less than 1.
(b) Assuming that the condition on q in part (a) holds true, determine the probability that
the population will eventually die out.
Exercise 3.4.4.∗ For a Galton-Watson Branching Process having X0 = n, n ∈ Z+ , calculate
the probability that the population will eventually die out when the number of offspring
(produced by any one individual) has pmf given by αm = (1 − β)m β, m = 0, 1, 2, . . . and
0 < β < 1.
1
Exercise 3.4.5. Consider a Galton-Watson Branching Process with X0 = 1 and µ < 1.
(a) How many individuals, on average, ever existed in this population?
(b) What would the answer to part (a) become if X0 = n, n ∈ Z+ ?
Exercise 3.4.6. For the Gambler’s Ruin Problem in Example 3.14, prove that {1, 2, . . . , N −
1} is a transient communication class.
Exercise 3.4.7.∗ Consider the following modification to the Gambler’s Ruin Problem. In-
stead of tossing one coin to determine whether the gambler wins or loses, two iid coins (each
with probability p of landing heads) are tossed. If both coins come up heads, the gambler
wins the game and receives $1. If both coins come up tails, he or she loses the game and pays
out $1. Otherwise, the gambler’s wealth remains the same. Let Xn denote the wealth of the
gambler after the nth play of the game. Assume that the gambler starts with $i and only quits
playing when he or she reaches either $0 or $N . Let Ui = P (gambler reaches $N |X0 = i).
(a) For i = 1, 2, . . . , N − 1, show that
Ui = aUi−1 + bUi+1
and find the values of a and b.
(b) Assuming that p ̸= 1/2, show that Ui = (1 − ri )/(1 − rN ), where r = [(1 − p)/p]2 .
Exercise 3.4.8. For the Gambler’s Ruin Problem in Example 3.14, let Mi denote the mean
number of games that must be played until the gambler either goes broke or reaches a fortune
of N units, given that he or she has an initial fortune of i units, i = 0, 1, . . . , N .
(a) Show that Mi satisfies
Mi = 1 + pMi+1 + qMi−1 , i = 1, 2, . . . , N − 1.
(b) Using M0 = MN = 0, solve the equations in part (a) to obtain
(
i(N − i) , if p = 12 ,
Mi = i N 1−(q/p) i
q−p
− q−p · 1−(q/p)N , if p ̸= 21 .