You are on page 1of 10

STAT 430 – Probability

Spring 2022
Homework 2
Due Friday, Jan 28th at 11:59pm
The following section and problem numbers refer to the book page images posted
on Canvas, which may differ from the book version you have. Be sure you are
working on the right problems. Answering the wrong homework problem will
lead to you receiving zero points for that problem.
You will upload your homework assignment to Gradescope. PDFs generated via
scan or via export from, e.g., LaTeX are preferred, although JPG files will also
be accepted. Late submissions are not allowed.

1. (10 points) Exercise 2.86


Solution 1. Recall that P(A ∩ B) = P(A) + P(B) − P(A ∪ B). Note that P(A ∪ B) ≤ 1,
and the equality happens when A ∪ B is the whole sample space. On the other hand,
we have P(A∪B) ≥ max{P(A), P(B)} = 0.8, and the equality is attained when B ⊂ A.
Thus, we have
P(A ∩ B) = 1.5 − P(A ∪ B) ∈ [0.5, 0.7],
and both the smallest and the largest values are attainable. This implies that:
(a) It is not possible that P(A ∩ B) = 0.1;
(b) The smallest possible value for P(A ∩ B) is 0.5;
(c) It is not possible that P(A ∩ B) = 0.77;
(d) The largest possible value for P(A ∩ B) is 0.7.
Solution 2. The sample space consists of four possible outcomes: A ∩ B, A ∩ B̄, Ā ∩
B, Ā ∩ B̄. We can calculate their probabilities in terms of P(A) = 0.8, P(B) = 0.7, and
P(A ∩ B) as follows:
P(A ∩ B) = P(A ∩ B).
P(A ∩ B̄) = P(A) − P(A ∩ B)
= 0.8 − P(A ∩ B).
P(Ā ∩ B) = P(B) − P(A ∩ B)
= 0.7 − P(A ∩ B).
P(Ā ∩ B̄) = 1 − P(A ∩ B) − P(A ∩ Ā) − P(Ā ∩ B)
= 1 + P(A ∩ B) − P(A) − P(B)
= −0.5 + P(A ∩ B).
Recall that all probabilities must fall within the interval [0, 1]. Thus, a value of P(A∩B)
is valid if and only if it results in the four quantities above being in the interval [0, 1].
We can use these facts to answer the assigned questions.
(a) No, P(A ∩ B) = 0.1 is not possible, because it implies that P(Ā ∩ B̄) = −0.4,
which does not fall within [0, 1].
(b) The smallest possible value for P(A ∩ B) is 0.5, which corresponds to the valid
probabilities P(A ∩ B̄) = 0.3, P(Ā ∩ B) = 0.2, and P(Ā ∩ B̄) = 0. Smaller values
of P(A ∩ B) result in negative values of P(Ā ∩ B̄), and are thus invalid.
(c) It is not possible P(A ∩ B) = 0.77, as it corresponds to P(Ā ∩ B) = −0.07, which
does not fall within [0, 1].
(d) The largest possible value of P(A ∩ B) is 0.7, corresponding to the valid proba-
bilities P(A ∩ B̄) = 0.1, P(Ā ∩ B) = 0, and P(Ā ∩ B̄) = 0.2. Larger values of
P(A ∩ B) result in negative values of P(Ā ∩ B), and are thus invalid.

2. (10 points) Exercise 2.88


Solution 1. First of all, we have P(A ∩ B) ≥ 0, and the quality happens when A
and B are disjoint. Meanwhile, P(A ∩ B) ≤ min{P(A), P(B)} = 0.3, and the equality
happens when B ⊂ A. From the Venn diagram, it is clear that any values from 0 to
0.3 is possible for P(A ∩ B). Thus, we have the following:

(a) It is possible that P(A ∩ B) = 0.1;


(b) The smallest possible value for P(A ∩ B) is zero;
(c) It is not possible that P(A ∩ B) = 0.7;
(d) The largest possible value for P(A ∩ B) is 0.3.

Solution 2.

(a) Yes. Because P(A ∩ B̄) = 0.5,P(Ā ∩ B) = 0.2 P(Ā ∩ B̄) = 0.2 (see Solution 2 of
problem 1), which are all between 0 and 1.
(b) 0. Because they can be disjoint.
(c) No. Because P(A ∩ B) cannot be larger than either P(A) or P(B).
(d) 0.3 , since P(A ∩ B) 6 P(B) = 0.3.

3. (10 points) Exercise 2.94


Solution. Let A and B denote the event that the smoke will be detected by device A
and B, respectively. P(A) = 0.95, P(B) = 0.9, P(A ∩ B) = 0.88

(a)

P(A ∪ B) = P(A) + P(B) − P(A ∩ B) (by the additive law)


= 0.95 + 0.9 − 0.88
= 0.97

(b)

P(Ā ∩ B̄) = 1 − P(A ∪ B) = 1 − 0.97 = 0.03 (by De Morgan’s laws)


Or more intuitively,

P(“smoke will be undetected”) = 1 − P(“smoke will be detected”)


= 1 − P(A ∪ B) = 1 − 0.97 = 0.03

4. (10 points) Exercise 2.96


Solution.

(a)

P(A ∪ B) = P(A) + P(B) − P(A ∩ B)


= P(A) + P(B) − P(A) × P(B) by independence
= 0.5 + 0.2 − 0.5 × 0.2
= 0.6

(b)

P(Ā ∩ B̄) = P(Ā) × P(B̄) by independence


= (1 − P(A)) × (1 − P(B))
= (1 − 0.5) × (1 − 0.2)
= 0.4

(c)

P(Ā ∪ B̄) = P(Ā) + P(B̄) − P(Ā ∩ B̄)


= (1 − P(A)) + (1 − P(B)) − P(Ā) × P(B̄) by independence
= 0.5 + 0.8 − 0.4
= 0.9

5. (10 points) Exercise 2.112


Solution.

(a) P(“undetected”) = 0.023 = 8 × 10−6 .


(b) P(“detected by all three”) = 0.983 ≈ 0.9412.

6. (10 points) Exercise 2.116


Solution. In order for the system to fail, all three lines must fail. Because they are
independent, the probability of this occurring is simply 0.013 = 0.000001. Thus, the
probability of not failing (the complement) is 1 − 0.000001 = 0.999999.

7. (10 points) At a gas station, 40% of the customers use regular gas (A1 ), 35% use mid-
grade gas (A2 ), and 25% use premium gas (A3 ). Of those using regular gas, only (30%)
fill their tanks (event B). Of those using mid-grade gas, 50% fill their tanks, whereas
of those using premium, 70% fill their tanks.
(a) What is the probability that the next customer requests mid-grade gas and fills
the tank?
(b) What is the probability that the next customer fills the tank?
(c) If the next customer fills the tank, what is the probability that she requests
premium gas?

Solution. We can use the facts

P(Ai ∩ B) = P(Ai ) × P(B|Ai ), and


P(Ai ∩ B̄) = P(Ai ) × P(B̄|Ai )

for i = 1, 2, 3 to construct the following table:

B B̄
A1 0.4 × 0.3 = 0.12 0.4 × 0.7 = 0.28
A2 0.35 × 0.5 = 0.175 0.35 × 0.5 = 0.175
A3 0.25 × 0.7 = 0.175 0.25 × 0.3 = 0.075

The entries in the table can then be used to solve the problems.

(a) From the table,

P(A2 ∩ B) = 0.175

(b) Adding entries in the table,

P(B) = P(A1 ∩ B) + P(A2 ∩ B) + P(A3 ∩ B)


= 0.12 + 0.175 + 0.175 = 0.47

(c) By definition of conditional probability,

P(A3 ∩ B)
P(A3 | B) =
P(B)
0.175
=
0.47
≈ 0.372

8. (10 points) For customers purchasing a full set of tires at a particular tire store, consider
events

A = {tires purchased were made in the United States}


B = {purchaser has tires balanced immediately}
C = {purchaser requests front-end alignment}.
Assume the following probabilities.

P (A) = .75 P (B|A) = .9 P (B|Ā) = .8


P (C|A ∩ B) = .8 P (C|A ∩ B̄) = .6 P (C|Ā ∩ B) = .7 P (C|Ā ∩ B̄) = .3 .

Find the probabilities P (A ∩ B ∩ C), P (B ∩ C), P (C), P (A|B ∩ C).


Solution. Note that it may be helpful to visualize this problem using a decision tree
with A and Ā being the node split, followed by B and C for levels 2 and 3, respectively.

(a)

P(A ∩ B ∩ C) = P(C | A ∩ B) × P(A ∩ B)


= P(C | A ∩ B) × P(B | A) × P(B)
= 0.8 × 0.9 × 0.75
= 0.54

(b)

P(B ∩ C) = P(Ā ∩ B ∩ C) + P(A ∩ B ∩ C)


= P(C | Ā ∩ B) × P(Ā ∩ B) + P(A ∩ B ∩ C)
= P(C | Ā ∩ B) × P(B | Ā) × P(Ā) + P(A ∩ B ∩ C)
= P(C | Ā ∩ B) × P(B | Ā) × (1 − P(A)) + P(A ∩ B ∩ C)
= 0.7 × 0.8 × (1 − 0.75) + 0.54
= 0.68

(c)

P(C) = P(C ∩ A ∩ B) + P(C ∩ Ā ∩ B) + P(C ∩ A ∩ B̄) + P(C ∩ Ā ∩ B̄)

Above, we decomposed the calculation into four summands. The first was already
computed in part (a). The other three can be calculated as

P(C ∩ Ā ∩ B) = P(C | Ā ∩ B) × P(Ā ∩ B)


= P(C | Ā ∩ B) × P(B | Ā) × P(Ā)
= P(C | Ā ∩ B) × P(B | Ā) × (1 − P(A))
= 0.7 × 0.8 × (1 − 0.75)
= 0.14.
P(C ∩ A ∩ B̄) = P(C | A ∩ B̄) × P(B̄ ∩ A)
= P(C | A ∩ B̄) × P(B̄ | A) × P(A)
= P(C | A ∩ B̄) × (1 − P(B | A)) × P(A)
= 0.6 × (1 − 0.9) × 0.75
= 0.045.
P(C ∩ Ā ∩ B̄) = P(C | Ā ∩ B̄) × P(B̄ ∩ Ā)
= P(C | Ā ∩ B̄) × P(B̄ | Ā) × P(Ā)
= P(C | A ∩ B̄) × (1 − P(B | Ā)) × (1 − P(A))
= 0.3 × (1 − 0.8) × (1 − 0.75)
= 0.015.

Thus, the final answer amounts to P(C) = 0.54 + 0.14 + 0.045 + 0.015 = 0.74.
(d)
P(A ∩ B ∩ C)
P(A | B ∩ C) =
P(B ∩ C)
0.54
=
0.68
≈ 0.794

9. (10 points) Exercise 2.142


Solution. Since we are spinning twice, our only options are to land on the same
position both times, or to land on two different positions. Thus, Y will either be 3
or 2. Note that the probability of landing on any particular region is 0.25. Thus, the
probability of landing on the same position twice is given by

P(AA ∪ BB ∪ CC ∪ DD) = P(AA) + P(BB) + P(CC) + P(DD) because the events are disjoint
(1)
= 0.252 + 0.252 + 0.252 + 0.252 assuming each spin is independent
(2)
= 0.25 (3)

Thus, P(Y = 3) = 0.25. It follows that P(Y = 2) = 1 − 0.25 = 0.75 as the only other
option. That is, P(Y = 0) = P(Y = 1) = P(Y = 4) = 0 because they are impossible in
only two spins.
10. (10 points) A random variable has pdf

c(2 − y) if 0 ≤ y ≤ 2
f (y) =
0 otherwise.

Find the value of c.

Solution. All pdfs must integrate to 1 over their domain. Thus,


Z 2
f (y)dy = 1
0
Z 2
⇒ c(2 − y)dy = 1
0
Z 2
⇒c (2 − y)dy = 1
0
22
 
⇒c 2(2) − =1
2
1
⇒c =
4−2
⇒c = 0.5

Be sure:
• To make this work your own even if you work with others! Do not blindly copy the
solutions; this is plagarism, and furthermore it will only hurt you at exam time.
• To put your name on all sheets, and to staple the sheets together, if you are turning
in multiple sheets.
• To properly label all plots (axes, main title, units, etc.) for full credit, if applicable,
and to include a printout of any code used in solutions, if applicable.

Optional problems. Optional problems are food for thought, covering related contents
we do not focus on, and/or slightly out of scope. You are welcome to openly discuss these
problems on Piazza at any time, or discuss them with the instructor during office hours.

O1. Exercise 2.109


Solution. The idea is to find the “worst case configuration” of those events. Consider
the following optimization problem:

min P(A ∩ B ∩ C)
subject to P(A) = P(B) = P(C) = p and all probabilistic constraints are satisfied.

Let the minimum value of the above optimization problem be f (p). Solving for f (p) ≥
0.95 gives an inequality for p. And this will tell us what does the final answer look
like.
Since p = P(A) ≥ P(A ∩ B ∩ C) and we want P(A ∩ B ∩ C) ≥ 0.95, we can without loss
of generality assume p ≥ 0.95. For the same reason, we can assume P(A ∩ B) ≥ 0.95.
Note that we can decompose P(A ∩ B ∩ C) = P(C | A ∩ B) · P(A ∩ B). To minimize
P(A ∩ B ∩ C), it suffices to first minimize P(C | A ∩ B) for every fixed configuration of
A ∩ B, and then minimize P(A ∩ B).
To minimize P(C | A ∩ B), we need the event C to intersect with A ∩ B as little as
possible. The probability of the complement of A ∩ B is 1 − P(A ∩ B) ≤ 0.05, whereas
P(C) ≥ 0.95. Thus, for the event C to intersect with A ∩ B as little as possible, the
event C need to occupy all of the complement of A ∩ B. Moreover, the event C need
to occupy part of A ∩ B, whose probability mass is P(A ∩ B ∩ C) = p − (1 − P(A ∩ B)).
This means that the minimal value of P(C | A ∩ B) is
p − 1 + P(A ∩ B)
.
P(A ∩ B)
Thus, the minimal value of P(A ∩ B ∩ C) is
p − 1 + P(A ∩ B).
Now it suffices to minimize P(A ∩ B). This is something we have done in Question 1
and 2. We have shown that the minimal value of P(A ∩ B) is
2p − 1,
which happens when A ∪ B is the whole sample space. Hence, the minimal value of
P(A ∩ B ∩ C) is
3p − 2.
Solving for 3p − 2 ≥ 0.95 gives
2.95
p≥ .
3
Note that p = 2.95
3
is attainable (we already had the construction above). Thus, the
smallest value for P(A) such that P(A ∩ B ∩ C) already exceeds 0.95 is
2.95
.
3
O2. Exercise 2.162.
Solution. For this question, we do not distinguish among cars of the same type, i.e.,
we assume all three sport cars are indistinguishable, etc. There are 7 ways to park
three sports cars adjacent to one another, and there are
 
9 9!
= = 3 · 8 · 7 · 10
3, 3, 3 3!3!3!
ways to park those cars. Thus, the probability that the three sport cars are parked
adjacent to one another is
7 1
= .
3 · 8 · 7 · 10 240
O3. Exercises 2.119, 2.138.
• Solution for Exercise 2.119. We only give the solution for Part (a). The
solution for Part (b) is similar. We want
P(See 3 before 7)
= P(See 3 in the first toss) · 1 + P(See 7 in the first toss) · 0
+ P(See neither 3 or 7 in the first toss)
× P(See 3 before 7 | See neither 3 or 7 in the first toss)
= P(See 3 in the first toss) + P(See neither 3 or 7 in the first toss)
× P(See 3 before 7 | See neither 3 or 7 in the first toss).
We have
2 2 6 28
P(See 3 in the first toss) = , P(See neither 3 or 7 in the first toss) = 1− − = .
36 36 36 36
If we see neither 3 or 7 in the first toss, we essentially “restart the experiment
from the second toss”. That is,
P(See 3 before 7 | See neither 3 or 7 in the first toss) = P(See 3 before 7).
Hence, we have
2 28
P(See 3 before 7) = + · P(See 3 before 7),
36 36
which implies
1
P(See 3 before 7) = .
4
Another solution. Let Xi be the sum of the two dices for the i-th toss, and let
T be the number of tosses that we first see a sum of either 3 or 7. We have

X X∞
P(See 3 before 7) = P(T = k)P(See 3 before 7 | T = k) = P(T = k)P(Xk = 3 | T = k).
k=1 k=1

We have
P(Xk = 3 ∩ T = k)
P(Xk = 3 | T = k) =
P(T = k)
P(Xk = 3)P(T = k | Xk = 3)
= .
P(Xk = 3)P(T = k | Xk = 3) + P(Xk = 7)P(T = k | Xk = 7)
Note that P(T = k | Xk = 7) = P(T = k | Xk = 3) by construction. Thus, we
have
P(Xk = 3) 1
P(Xk = 3 | T = k) = = .
P(Xk = 3) + P(Xk = 7) 4
This implies

X 1 1
P(See 3 before 7) = P(T = k) · = .
k=1
4 4
• Solution for Exercise 2.138. The probability that the player wins after the
first toss is
6+2 8
= .
36 36
The probability that the player loses after the first toss is
2+2+2 6
= .
36 36
The probability that the player continues after the first toss is
8+6 22
1− = .
36 36
Conditional on that the player continues after the first toss, we can use the ar-
guments in Exercise 2.138 to compute the probability of wining for each value of
the first toss. Summarizing the above will give the answer of this exercise.

You might also like