IE5004, Homework 8
Solution to 7.1.
Let X be the winnings and Y be the outcome of the die. Let H = {the coin lands heads} and
T = {the coin lands tails}. Then,
(
2Y if H occurs,
X=
Y /2 if T occurs,
so that
E[X] = E[2Y |H]P[H] + E[Y /2|T ]P[T ].
Because the die and the coin are flippled independently, E[2Y |H] = E[2Y ] and E[Y /2|T ] = E[Y /2].
Using the fact that E[Y ] = (1 + 2 + · · · + 6)/6 = 3.5, we have
E[X] = E[2Y ]P[H] + E[Y /2]P[T ] = E[Y ] × 2 × 0.5 + E[Y ] × 0.5 × 0.5 = 3.5 + 0.875 = 4.375.
Solution to 7.21.
For all k = 1, . . . , 365, assume the probability that a person’s birthday is on day k is 1/365.
(a) Let Ik be the indicator variable such that
(
1 day k is the birthday of exactly three people,
Ik =
0 otherwise.
Then,
100 1 3 1 100−3
E[Ik ] = P[day k is the birthday of exactly three people] = 1− .
3 365 365
Let N = I1 + · · · + I365 be the number of days of the year that are birthdays of exactly three
people. It follows that
365
X 100 1 3 1 97
E[N ] = E[Ik ] = 365 1− .
3 365 365
k=1
(b) For k = 1, . . . , 365, let Xk be the indicator variable such that
(
1 if day k is someone’s birthday,
Xk =
0 otherwise.
Then,
364 100
E[Xk ] = P[day k is someone’s birthday] = 1 − P[day k is not a birthday] = 1 − .
365
Let M = X1 + · · · + X365 be the number of distinct birthdays. We have
365
X 364 100
E[M ] = E[Xk ] = 365 1 − .
365
k=1
1
Solution to 7.58.
Let X denote the number of flips required.
(a) Let H = {the first flip is heads} and T = {the first flip is tails}. If the first flip is heads, then
X is the number of flips until the first tails appears. In other words, given H occurs, X − 1
is a geometric random variable with probability 1 − p, i.e.,
P[X − 1 = k|H] = pk−1 (1 − p) for k = 1, 2, . . . .
This geometric random variable has expected value
1
E[X − 1|H] = .
1−p
Similarly, if the first flip is tails, X is the number of flips until the first heads appears. Then,
P[X − 1 = k|T ] = (1 − p)k−1 p for k = 1, 2, . . .
and
1
E[X − 1|T ] = .
p
It follows that
E[X] = E[X|H]P[H] + E[X|T ]P[T ]
1 1
= 1+ p+ 1+ (1 − p)
1−p p
p 1−p
=1+ + .
1−p p
(b) Let A = {the last flip is heads}. If A occurs and we flip the coin k times until both heads and
tails have appeared, the first k − 1 flips must be all tails. Then,
∞
X ∞
X
P[A] = P[A, X = k] = (1 − p)k−1 p = 1 − p.
k=2 k=2
In fact, if you are able to see A = {the last flip is heads} = {the first flip is tails}, you can
find the answer immediately.
Solution to 8.7.
Let Xk be the lifetime of the kth bulb. Then, E[Xk ] = 5 and V (Xk ) = 25. Using the central
limit theorem, we have
100
X P100
k=1 (Xk − 5) 525 − 500
P Xk > 525 = P √ > √ ≈ 1 − Φ(0.5) = 1 − 0.6915 = 0.3085.
k=1
100 · 25 100 · 25
Solution to 8.8.
2
Let Yk be the replacing time for the kth bulb. Then, E[Yk ] = 0.25 and V (Yk ) = 0.52 /12 = 1/48.
We would compute the probability
100
X 99
X
P Xk + Yk < 550 .
k=1 k=1
P100 P99
Since k=1 Xk + k=1 Yk has expected value 100 × 5 + 99 × 0.25 = 524.75 and variance 25 × 100 +
99/48 = 2502.06, it follows that the desired probability is approximately equal to
P100 P99
k=1 Xk + k=1 Yk − 524.75 550 − 524.75
P √ < √ ≈ Φ(0.505) = 0.6950.
2502.06 2502.06
Solution to 8.10.
Let Wn be the total weight of n cars and X be the amount of weight that the bridge can
withstand. Then, Wn − X is normally distributed with expected value 3n − 400 and variance
0.09n + 1600. Hence, the probability of structural damage is
Wn − X − (3n − 400) 400 − 3n 400 − 3n
P[Wn − X > 0] = P √ >√ ≈1−Φ √ .
0.09n + 1600 0.09n + 1600 0.09n + 1600
Since Φ(1.28) = 0.8997, to make P[Wn − X > 0] > 0.1, we can set
400 − 3n
√ < 1.28.
0.09n + 1600
The smallest integer that satisfies the above inequality is n = 117.
