Professional Documents
Culture Documents
Spring 2019
1.
Solution.
Solution.
ZZ Z 1 Z x
E(XY ) = xyf (x, y)dxdy = xy · 6y dy dx
0 0
Z 1 h
y=x i
= 2xy 3 y=0 dx
0
Z 1 1
4 2 5 2
= 2x dx = x =
0 5 0 5
2
So
2 1 3 1
Cov(X, Y ) = E(XY ) − E(X)E(Y ) = − · =
5 2 4 40
Thus, X and Y are dependent since Cov(X, Y ) 6= 0.
2.
(1) Suppose that X has density function given by
(
2x, 0 ≤ x ≤ 1,
f (x) =
0, elsewhere.
Find the probability density function for Y = eX .
3.
(a) Compute the moment generating function mX (t) for t > −1.
(b) Use part (a) to compute E(e2X ), E(X) and E(X 3 ).
1
(b) E(e2X ) = mX (2) = 1+2
= 13 . Since
(3)
m0X (t) = −(1 + t)−2 , m00X (t) = 2(1 + t)−3 , mX (t) = −6(1 + t)−4 ,
(3)
we get E(X) = m0X (0) = −(1 + 0)−2 = −1 and E(X 3 ) = mX (0) =
−6(1 + 0)−4 = −6.
(2) (a) Suppose that the random variable Y has Bernoulli distribution
with rate p ∈ (0, 1), that is, its probability distribution is
(
p, y = 1,
p(y) = P (Y = y) =
1 − p, y = 0.
Solution. (a)
X
mY (t) = E(etY ) = ety p(y) = et·1 p(1) + et·0 p(0) = pet + (1 − p).
all y
which is exactly the mgf of Binomial distribution B(n, p), that is,
U ∼ B(n, p).
(3) Find the distributions of the random variables that have the MGFs:
5
(a) mX (t) = [(1/3)et + (2/3)] .
et
(b) mY (t) = .
2 − et
t
(c) mZ (t) = e2(e −1) .
2
(d) mU (t) = e−2t+2t .
e2t − et
(e) mV (t) = .
t
1
(f) mW (t) = .
(1 − 3t)5
Solution.
5 5
(a) mX (t) = [(1/3)et + (2/3)] = [(1/3)et + (1 − 1/3)] , so X ∼ B(5, 13 ).
1 t
et e
(b) mY (t) = = 2
, so Y ∼ Geo( 12 ).
2−e t 1 − (1 − 21 )et
t
(c) mZ (t) = e2(e −1) , so Z ∼ P oisson(2).
2 1 2 2
(d) mU (t) = e−2t+2t = e−2t+ 2 ·2 t , so U ∼ N (−2, 22 ).
7
4.
So Z ∞
1 (u−800)2
P (U > 1050) = √ e− 2·2502 du ≈ 15.87%
1050 250 2π
(a) Find the probability that the mean of Population 1 is between 294
and 306.
(b) Find the probability that the mean of Population 1 is greater than
the mean of Population 2 by more than 20.
Solution. (a) Let {Xi }1≤i≤n be the samples of population 1, then the
sample mean
(3) Suppose that from past data a professor knows that the test score of
a typical student taking their final examination is a normal random
variable with mean 73 and standard deviation 10.
(a) If 4 students are selected at random, what is the probability that
their sample average grade will be within 5 of 73?
(b) What is the minimum number of students that need to take the
examination to ensure, with probability at least 99.7%, that the
class average would be within 5 of 73? (Hint: Using the fact that
if Y ∼ N (µ, σ 2 ), then P (|Y − µ| ≤ 3σ) ≈ 99.7%).
Solution. (a) Here n = 4 and let Xi be the test score of the i-th student
for i = 1, . . . , n. Since each Xi ∼ N (73, 102 ), the sample mean
X n ∼ N (73, 102 /4) = N (73, 52 ).
Then the probability that their sample average grade will be within 5
of 73 is
Z 78
1 (x−73)2
P (|X n −73| ≤ 5) = P (68 ≤ X n ≤ 78) = √ e− 2·52 dx ≈ 68.27%
68 5 2π
10
5.
(1) The service times for customers coming through a checkout counter in
a retail store are independent random variables with mean 1.5 minutes
and variance 1.0. Use the Central Limit Theorem to approximate the
probability that 100 customers can be served in less than 2 hours of
total service time.
Solution. Here n = 100 and let Yi be the service time for the i-th cus-
tomer, for i = 1, 2, . . . , n. Since they are independent, by the Central
Limit Theorem,
Y 100 − 1.5
U= √ = 10Y 100 − 15 ≈ N (0, 12 ),
1/ 100
and thus the probability that 100 customers can be served in less than
2 hours is
So the probability that there are not enough seats will be available is
372 − 360
P (Y > 372) = P U > = P (U > 2)
6
Z ∞
1 2
= √ e−x /2 dx ≈ 2.28%
2 2π
(3) The lifetime of light bulbs from a certain factory has an exponential
distribution with mean 18 months.
(a) 81 light bulbs are purchased from this company. Use the Central
Limit theorem to approximate the probability that the average
lifetime of the 81 light bulbs is between 14 and 16 months.
(b) Now suppose we have 100 light bulbs. Use the Central Limit
Theorem to approximate the probability that the total lifetime
of these 100 bulbs (that is, the sum of their lifetimes) does not
exceed 2000 months.
Solution. (a) Here n = 81 and let Yi be the lifetime of the i-th light
bulb, for i = 1, 2, . . . , n. Each Yi ∼ exp(18), and hence E(Yi ) = 18
and V (Yi ) = 182 . Since they are independent, by the Central Limit
Theorem,
Y 81 − 18 Y 81 − 18
U= √ = ≈ N (0, 12 ),
18/ 81 2
and thus the probability that the average lifetime of the 81 light bulbs
is between 14 and 16 months is
14 − 18 16 − 18
P (14 ≤ Y 81 ≤ 16) = P ≤U ≤
2 2
= P (−2 ≤ U ≤ −1)
Z −1
1 2
= √ e−x /2 dx ≈ 13.59%
−2 2π
13
(b) Now n = 100 and let Yi be the lifetime of the i-th light bulb, for
i = 1, 2, . . . , n. Each Yi ∼ exp(18), and hence E(Yi ) = 18 and V (Yi ) =
182 . Since they are independent, by the Central Limit Theorem,
Y 100 − 18 Y 100 − 18
U= √ = ≈ N (0, 12 ),
18/ 100 1.8
and thus the probability that the total lifetime of these 100 bulbs does
not exceed 2000 months is