3F Probability

Let X be a random variable taking non-negative integer values and let Y be a

random variable taking real values.
(a) Define the probability-generating function GX (s). Calculate it explicitly for a
Poisson random variable with mean λ > 0.

GX (s) = E sX = ∞ n

P
n=0 s P(X = n). For X ∼Po(λ),

∞ ∞
X λn X (sλ)n
GX (s) = sn e−λ = e−λ = eλ(s−1) .
n! n!
n=0 n=0
[3]
(b) Define the moment-generating function MY (t). Calculate it explicitly for a
normal random variable N(0, 1).

MY (t) = E etX . For Y ∼ N(0, 1),

Z ∞
1 2
MY (t) = √ etx e−x /2 dx
2π −∞
t2 /2 Z ∞
e 2 2
=√ e−(x−t) /2 dx = et /2 ,
2π −∞
√ 2
as the last integral gives 2π (since e−(x−t) /2 2π is the PDF for
√

[3] N(t, 1)).

(c) By considering

a random sum of independent copies of Y , prove that, for general
X and Y , GX MY (t) is the moment-generating function of some random variable.

Let X, Y1 , Y2 , . . . be independent reandom variables with Yn ∼ Y for all

XX
n. Define the random variable Z = Yn .
n=1

We then have (for all t)

∞
X
tZ
P(X = n)E etZ X = n



MZ (t) = E e =
n=0
X∞  
= P(X = n)E et(Y1 +···+Yn )
n=0
X∞ n
Y
E etYi


= P(X = n) since the Yi are indep
n=0 i=1
∞
X
n
P(X = n)E etY


= = GX MY (t) .
n=0
[4]
[10]
 9F Probability
(a) Let B1 , . . . , Bn be pairwise disjoint events such that their union B1 ∪B2 ∪. . .∪Bn
gives the whole set of outcomes, with P(Bi ) > 0 for 1 6 i 6 n. Prove that for any event
A with P(A) > 0 and for any i

P(A|Bi )P(Bi )
P(Bi |A) = P .
P(A|Bj )P(Bj )
16j6n

By definition of conditional probability, for each i

P(Bi ∩ A) P(A|Bi )P(Bi )

P(Bi |A) = = . (1)
P(A) P(A)

Since the Bj are pairwise disjoint and cover all outcomes,

[  X
P(A) = P (A ∩ Bj ) = P(A ∩ Bj ).
j j

The RHS can be rewritten as

X
P(A|Bj )P(Bj ),
j

[5] and subbing into (1) for P(A) gives the result.
(b) A prince is equally likely to sleep on any number of mattresses from six to eight;
on half the nights a pea is placed beneath the lowest mattress. With only six mattresses
his sleep is always disturbed by the presence of a pea; with seven a pea, if present, is
unnoticed in one night out of five; and with eight his sleep is undisturbed despite an
offending pea in two nights out of five.
What is the probability that, on a given night, the prince’s sleep was undisturbed?

Let U be the event that he slept undisturbed, P be the event that a pea
is placed (with complement P c ), and N be the number of mattresses.
Assuming P and N are indep, we have
8
X
P(U ) = P(U ∩ P c ) + P(U ∩ P ∩ {N = n})
n=6
8
1 X
= + P(U |P ∩ {N = n}) × P(P ) × P(N = n)
2
n=6
8  
1 Xn−6 1 1 1 1 1 2 3
= + × × = + 0+ + = .
2 5 2 3 2 6 5 5 5
n=6
[7]
 On the morning of his wedding day, he announces that he has just spent the most
peaceful and undisturbed of nights. What is the expected number of mattresses on which
he slept the previous night?

(c) For each n ∈ {6, 7, 8} we have

P(U |N = n)P(N = n) 5
P(N = n|U ) = = × P(U |N = n).
P(U ) 9
And
P(U |N = n) = P(P )P(U |P ∩ {N = n}) + P(P c )P(U |P c ∩ {N = n})
n−6 1 n−1
= + = .
10 2 10
So
n−1
P(N = n|U ) =
18
and hence
8
X
E(N |U ) = nP(N = n|U )
n=6
8
X n(n − 1) 1 128 64
= = (30 + 42 + 56) = = .
18 18 18 9
n=6
[8]
[20]

12F Probability
A circular island has a volcano at its central point. During an eruption, lava flows
from the mouth of the volcano and covers a sector with random angle Φ (measured in
radians), whose line of symmetry makes a random angle Θ with some fixed compass
bearing.
The variables Θ and Φ are independent. The probability density function of Θ is
constant on (0, 2π) and the probability density function of Φ is of the form A(π − φ/2)
where 0 < φ < 2π, and A is a constant.
(a) Find the value of A. Calculate the expected value and the variance of the sector
angle Φ. Explain briefly how you would simulate the random variable Φ using a uniformly
distributed random variable U .
Z 2π
By normalisation of the PDF, A(π − φ/2) dφ = 1
0
2π 1
Hence 1 = A πφ − φ2 /4 0 = π 2 A. Hence A = 2 .

π
Then
Z 2π
2π 2π
Aφ(π − φ/2) dφ = A πφ2 /2 − φ3 /6 0 = A(2π 3 − 4π 3 /3) =

E(Φ) = .
0 3
 and
Z 2π 2π
2
Aφ2 (π − φ/2) dφ = A πφ3 /3 − φ4 /8 0

E(Φ ) =
0
2π 2
= A(8π 4 /3 − 2π 4 ) =
3
so
2π 2 4π 2 2π 2
Var(Φ) = E(Φ2 ) − E(Φ)2 = − = .
3 9 9
Φ has distribution function

0
 for φ 6 0
FΦ (φ) = P(Φ 6 φ) = φ/π − φ2 /4π 2 for φ ∈ [0, 2π]

1 for φ > 2π.


For x ∈ (0, 1) the inverse FΦ−1 (x) is a solution to

φ2 φ
2
− +x=0
4π π
and is therefore
p
1/π ± 1/π 2 − x/π 2 √

= 2π 1 ± 1 − x .
1/2π 2

Since Φ 6 2π we must have the minus in ±. Hence if U ∼ U(0, 1) then

√
FΦ−1 (U ) = 2π 1 − 1 − U

[7] simulates Φ.
(b) H1 and H2 are two houses on the island which are collinear with the mouth of
the volcano, but on different sides of it. Find
(i) the probability that H1 is hit by the lava;

WLOG let H1 and H2 be at angles θ = 0 and π from the volcano mouth,

w.r.t the given fixed bearing. Then
Z 2π Z 2π
P(H1 hit) = fΘ (θ)fΦ (φ)1{θ<φ/2}∪{2π−φ/2<θ} dθ dφ
0 0
Z 2π
φ 1
= fΦ (φ) dφ = E(Φ)/2π = .
0 2π 3
[4]
(ii) the probability that both H1 and H2 are hit by the lava;
 Z 2π Z 2π
P(Both hit) = fΘ (θ)fΦ (φ)1{π−φ/2<θ<φ/2}∪{2π−φ/2<θ<π+φ/2} dθ dφ
0 0
2π
2(φ − π)
Z
= fΦ (φ) dφ
π 2π
Z 2π
1
= 3 (φ − π)(π − φ/2) dφ
π π
Z π
1
= 3 ψ(π − ψ) dψ by substitution
2π 0
1  π 1 π3 1
= 3 πψ 2 /2 − ψ 3 /3 0 = 3 × = .
2π 2π 6 12
[5]
(iii) the probability that H2 is not hit by the lava given that H1 is hit.

P(H2 not hit |H1 hit) = 1 − P(H2 hit |H1 hit)

P(Both hit) 1/12 3
=1− =1− = .
P(H1 hit) 1/3 4
[4]
[20]