Q1 7 points

It is a relatively rare event that a new television show becomes a long-term success. A new television
show that is introduced during the regular season has a 10% chance of becoming a long-term
success. A new television show that is introduced as a mid-season replacement has only a 5% chance
of becoming a long-term success. Approximately 60% of all new television shows are introduced
during the regular season. What is the probability that a randomly selected new television show will
become a long-term success?

Answer:
P(Long-term success) = P(Long-term success | Regular-season show) x P(Regular-season show) +
P(Long-term success | Middle-season show) x P(Middle-season show) = (0.10) x (0.60) + (0.05) x
(0.40) = 0.08.

Q2 6 points

In the Filatoi Riuniti case we minimize the total cost. The capacity of Ambrosi was 2500, with a
shadow price of -4.15 (allowable increase of 693.6 and allowable decrease of 1832). Which of the
following statements are true (it can be more than 1)?

A The shadow price is negative so it is bad to increase the capacity

b The shadow price is negative so it is good to increase the capacity
c If we increase the capacity by 700, it will decrease the objective function by 700*4.15
d If we increase the capacity by 700, it will inrease the objective function by 700*4.15
e If we decrease the capacity by 700, it will decrease the objective function by 700*4.15
f If we decrease the capacity by 700, it will increase the objective function by 700*4.15

Q3 16 points

The average winter temperature in Stockholm is normally distributed with mean, m=0.5 degrees and
standard deviation, s = 1.45 degrees. In Gothenburg the average winter temperature is normally
distributed with mean, m=1.5 degrees and standard deviation, s = 2.45 degrees.

a) What is the probability that the average temperature in Stockholm this coming winter will be
above freezing (0 degrees)?

Answer:
Let X be the temperature in Stockholm. P(X>0) = 1-NORM.DIST(0,0.5,1.45,1) = 0.635

6 points – correct answer

3 points – forgot 1-…
2 points – correct start but wrong answer
b) Assume that the temperatures in Stockholm and Gothenburg are independent. What is the
probability that the average winter temperature this coming winter is higher in Stockholm than in
Gothenburg?

Answer:
Let Y be the temperature in Gothenburg and D=X-Y the temperature difference. The average winter
temperature this coming winter is higher in Stockholm than in Gothenburg if D>0.

E(D) = 0.5-1.5 = -1
Var(D) = (1)^2Var(X)+(-1)^2Var(Y) = 1.45^2+2.45^2=8.105
SD(D) = sqrt(8.105) = 2.847

P(D>0) = 1-P(D<=0) = =1-NORM.DIST(0,-1,2.847,1) = 0.362

6 points – correct answer

2 points - correct use of formulas, correct part answer but incorrect final answer
2 points - P(E(temp in sthlm) > temp I Gothenburg)
1 points – formula but rest wrong

c) What do you think of the assumption that the temperatures in Stockholm and Gothenburg are
independent? Motivate your answer!

Answer:
It is likely that they are not independent meaning that the temperature in the two cities during a
winter is correlated. The opposite answer is also ok with good argumentation.

4 points – correct answer

2 point - some explanation of independence related to the question/case but with flawed
arguments
1 point - some explanation of independent but not relevant to the question/case

Q4 10 points

You want to invest in two equity funds: Sweden & Global. Let X denote the annual expected return
of Sweden and Y the annual expected return of Global. We assume that both X and Y follow normal
distributions with E(X)=E(Y)=m and Var(X)=Var(Y)=s^2. You decide to invest 50% in Sweden and 50%
in Global.

(a) What is the standard deviation of the annual return of the total investment if corr(X,Y)=-1?

Answer: 0 since they are perfectly negatively correlated

10 points – correct answer
6 points – All correct up to final answer, which was incorrect
5 points – correct but no calculations
1 points – used corr=1 and answered var instead of std
1 points – formula only

Q5 20 points

Health & Comfort, Incorporated (H&C) specializes in the production of hypoallergenic pillow and
mattress covers for asthma and allergy sufferers. The covers are made out of a unique protective
fabric. The manufacturing process consists mainly of two operations: cutting and sewing, which are
performed by two separate departments. The table below shows the resource utilization and weekly
resource availability for this problem. H&C is one of the largest producers of hypoallergenic bedding,
and it controls a large share of the market. As a consequence, the production level of each product
at H&C influences the market price of the product, which in turn influences their unit earnings
contributions. Econometric research indicates that this dependence can be represented as

EP = 506 - 0.75P

for pillow covers, and

EM = 653.5 - M

Pillow Cover Mattress Cover Availability per Week

Fabric (square feet) 8.75 60.5 40,000

Cutting (hours) 0.1 0.2 150
Sewing (hours) 0.4 0.3 500

for mattress covers (here P and M are the production levels of pillow and mattress covers at H&C,
and EP and EM are the corresponding unit contributions to earnings in dollars per unit). H&C is
seeking to maximize its earnings by selecting appropriate production levels of its products.

a) Formulate the objective function in terms of the variables P and M

Answer:
(506 – 0.75P)P + (653.5 – M)M

6 points – all correct

b) Formulate the constraints

Answer:

Fabric: 8.75P + 60.5M <= 40,000,

Cutting: 0.1P + 0.2M <= 150,

Sewing: 0.4P + 0.3M <= 500,

Pillow price: 506 – 0.75P >= 0,

Mattress cover price: 653.5 – M >= 0,

Non-negativity: P, M >= 0

12 points for all constraints (2 per constraint)

1 point per constraint with incorrect sign

c) How would you classify the new problem? Linear, nonlinear, discrete, continuous, integer,
maximization, minimization, etc. Motivate why!

Answer: Nonlinear, maximization

Q6 20 points

The file hprice.xlsx show housing prices from 506 communities in the Boston area. Description of the
variables can be found in the tab ” Variables”.

a) Run a regression to investigate how housing prices depend on pollution, crime, rooms and
disctance. Are all the coefficients of the sign that you would have expected? Explain briefly why! Are
they all significant?

Answer: Yes (sign and significant)

7 points – all correct

3 points – right interpretations and explanations but not correct regarding significance

b) Predict the average price of a house with nox=3, crime=4, rooms=5 and distance=6. Show your
calculations!

Answer:
From the regression:

Price_hat = -9806.39-213.53*3-2381.24*4+7843.90*5-786.92*6 = 16693.6751

7 points

c) You suspect that the effect of rooms on price in not linear. How would you go about testing that?
What do you find?

Answer:
Add squared rooms  significant

6 points

Q7 20 points

The attached file, “FILATOIR_exam”, contains the optimal solution of the Filatoir case.

a) The controller also tells you that the total capacity at Filatoi will be 3000 lower in the
upcoming year. Will this change the optimal solution? Why?

ANSWER: Yes, because the allowable decrease for the shadow price for filatoi capacity is 2290

6 points

b) For political reasons there is a need to purchase at least the same quantity (kg) from
Ambrosi as from De Blasi. Formulate this constraint. Remember that Xij is the amount
bought from supplier i of yarn size j

ANSWER: a) X11 + X12 + X13 + X14 => X41 + X42 + X43 + X44

7 points

c) For fine yarn, Filatoi also needs to purchase at least 50% of the quantity (kg) they buy from
Ambrosi from De Blasi. Formulate this constraint

ANSWER: X12*0.5 <= X42

7 points