CASE 4.2 A MERICAN O FFICE S YSTEMS , I NC .
11
A merican Office Systems, Inc., was established by
the late R. J. Miller, Sr., in 1939. It started as an
office supply story in Mountain View, California, and
expanded slowly over the years into the manufacture
of small office equipment, overhead projectors, and
bookkeeping machines. In the 1950s, computers
started eroding its market for bookkeeping machines,
so the company diversified into the copy machine
market. However, it never captured a large market
share because bigger firms such as Xerox, Canon,
Sharp, and A. B. Dick were so firmly entrenched.
A few years ago, American Office Systems’
engineering staff developed an adapter that links a
standard copy machine to personal computers,
allowing a copy machine to be used as a laser printer,
scanner, and fax. The adapters show great promise
for both home and office use. However, the company
is not well known by either the financial community
or the copy machine market, principally due to its
small size and rather lackluster record, so it could
secure only $15 million in initial financial backing for
the adapters. The $15 million was used to finance
the construction of a small production facility and of
administrative offices in 1994, and in 1995 produc-
tion and sales began. Two versions of the adapter
exist, one for IBM-compatible computers and one for
Macintosh computers. The former sells for $175 and
the latter for $200.
At the beginning of December 1995, Dr. R. J.
Miller, II, President, convened a meeting about the
coming year’s plans for the adapters. Rob Olsen,Vice
President of Production, argued that production
facilities should be expanded: “Until we have sufficient
capacity to produce the adapters,” he said,“there is no
use advertising.” Sue Williams, Director of Marketing,
replied,“On the contrary, without any demand for the
adapters, there is no reason to produce them. We
need to focus on advertising first.” J.T. Howell, the
Comptroller, pointed out that Olsen and Williams
were talking about the situation as if it only involved a
decision between production and marketing: “Yes,
funds need to be allocated between production and
advertising. However, more important than both is
the cash flow difficulty that the company has been
11
This case was written by William D. Whisler, California State
University, Hayward.
210 Chapter 4 Linear Programming Models
experiencing. As you know, it was only yesterday that,
finally, I was able to secure a $750,000 line of credit
for the coming year from Citibank. I might add that it
is at a very favorable interest rate of 16%. This will
partially solve our cash flow problems and it will have
a big effect on both production and advertising deci-
sions. In addition, there are financial and accounting
factors that must be allowed for in any decision about
the adapters.” Olsen interjected, “Wow, this is more
complicated than I anticipated originally. Before we
make a decision, I think we ought to use some
modern management science techniques to be sure
that all the relevant factors are considered. Last week
I hired Carlos Garcia from Stanford. He has a Master’s
Degree in Operations Research. I think this would be
a good project for him.” However, Williams said that
she thinks that an executive, judgmental decision
would be much better.“Let’s not get carried away with
any of the quantitative mumbo-jumbo that Rob is
always suggesting. Besides, his studies always take too
much time and are so technical that no one can
understand them. We need a decision by the end of
next week.” After listening to the discussion, Miller
decided to appoint an executive action team to study
the problem and make a recommendation at next
week’s meeting.“Rob and Sue, I want both of you to
document your arguments in more detail. J.T., be more
precise with your comments about the cash flow,
accounting, and financial problems. And, by the way
Rob, have Carlos look into a model to see if it might
produce some insights.”
Most of the $15 million initial financing was used
to build a five-story building in Mountain View, south
of San Francisco. Although currently only about 90%
complete, it is being used. The first floor contains the
production and shipping facilities plus a small storage
area. A larger warehouse, already owned by the
company, is located across the street. The other four
floors of the building are for the engineering depart-
ment (second floor), a research lab (third floor), and
administration (top two floors). The production
facility operates two shifts per day and has a produc-
tion capacity of 30 IBM adapters and 10 Macintosh
adapters per hour. Olsen uses 20 production days
per month in his planning. Usually there are a few
more, but these are reserved for maintenance and
repairs. The last stage of the initial construction will
be finished by the beginning of the fourth quarter,
making the building 100% finished. This will increase
the production capacity rates by 10%.
Howell normally does the company’s financial
planning monthly, and he assumes that cash flows as-
sociated with all current operating expenses, sales
revenues (taking collections into account), advertising
costs, loans from the line of credit, investments of
excess cash in short-term government securities, and
so forth, occur at the end of the corresponding
month. Because he needs information for the meeting
next week, however, he decides to do a rough plan on
a quarterly basis. This means that all the just men-
tioned cash flows, and so on, will be assumed to occur
at the end of the quarter. After the meeting, when
more time is available, the plan will be expanded to
a monthly basis. To get started, one of his senior
financial analysts prepares the list of quarterly fixed
operating expenses shown in Table 4.16. In addition,
the accounting department calculates that the variable
costs of the adapters are $100 each for the IBM
version and $110 each for the Macintosh version.
Table 4.16 Quarterly Fixed Operating Expenses
Expense Cost
Administrative expense $1,500,000
Fixed manufacturing costs 750,000
Sales agents’ salaries 750,000
Depreciation 100,000
At present, American Office Systems is
experiencing a cash flow squeeze due to the large
cash requirements of the startup of the adapter
production, advertising, and sales costs. If excess cash
is available in any quarter, however, Howell says that
the company policy is to invest it in short-term
government securities, such as treasury bills. He
estimates that during the coming year these
investments will yield a return of 6%.
Olsen asks Garcia to look into the production
and inventory aspects of the situation first, because
this area was his specialty at Stanford. Then he says
that he wants him to think about a programming
model that might integrate all components of the
problem—production, sales, advertising, inventory,
accounting, and finance. A mixed-integer programming
model appears to be the most appropriate; however,
he asks Garcia to use linear programming as an
approximation due to the time limitations and
Williams’s concern about his ideas always being too
technical. “There will be more time after next week’s
meeting to refine the model,” he says.
After discussions with Olsen and Williams,
Garcia feels that something needs to be done to help
the company handle the uncertainty surrounding
future sales of the adapters. He points out that it is
impossible to guarantee that the company will never
be out of stock. However, it is possible to decrease
shortages so that any difficulties associated with
them would be small and they would not cause
major disruptions or additional management prob-
lems, such as excess time and cost spent expediting
orders, and so forth. Thus, Garcia formulates an
inventory model. To be able to solve the model, he
has to check the inventory levels of the adapters
currently on hand in the warehouse. From these
quantities, he calculates that there will be 10,000 IBM
and 5,000 Macintosh adapters on hand at the begin-
ning of 1996. Based on the results of the model, he
recommends that a simple rule of thumb be used:
production plus the end-of-period inventory for the
adapters should be at least 10% larger than the
estimated sales for the next period. This would be
a safety cushion to help prevent shortages of the
adapters. In addition, to provide a smooth transition
to 1997, the inventory level plus production at the
end of the fourth quarter of 1996 should be at least
twice the maximum expected sales for that quarter.
Garcia says that using these rules of thumb will
minimize annual inventory costs. When explaining
the inventory model to Olsen, Garcia emphasizes
the importance of including inventory carrying costs
as part of any analysis, even though such costs
frequently are not out-of-pocket. He says that his
analysis of data provided by the accounting depart-
ment yielded a 1% per month inventory carry cost,
and this is what he used in his model.
Sales during the first year (1995) for the adapters
are shown in Table 4.17. Next year’s sales are
Case 4.2 American Office Systems, Inc. 211
uncertain. One reason for the uncertainty is that
they depend on the advertising. To begin the analysis,
Williams asks her marketing research analyst, Debra
Lu, to estimate the maximum sales levels for the
coming four quarters if no advertising is done. Since
last year’s sales of both models showed a steady
increase throughout the year, Lu projects a
continuation of the trend. She forecasts that the
company will be able to sell any number of adapters
up to the maximum expected sales amounts shown
in Table 4.17.
Table 4.17 1995 Adapter Sales and Maximum
Expected 1996 Sales
1996 Maximum
1995 Sales Expected Sales
IBM Macintosh IBM Macintosh
Quarter Adapters Adapters Adapters Adapters
1 5,000 1,000 9,000 1,800
2 6,000 1,200 10,000 2,000
3 7,000 1,400 11,000 2,200
4 8,000 1,600 12,000 2,400
Miller suggests that advertising in magazines
such as PC World and Home Office will increase
consumer awareness of both the company and
adapters. The next day, Williams has a meeting with
several staff members of a San Francisco advertising
agency. They show her recommendations for two
types of ads (one for the IBM adapters and one for
the Macintosh adapters), give her cost information,
and the estimated effectiveness of an advertising
campaign. Armed with this information and some
data from Lu, Williams prepares a brief report for
Miller setting out her reasons for thinking that each
$10 spent on advertising will sell an additional IBM
adapter; the same relationship holds true for the
Macintosh adapter.
Based on an analysis of 1995 sales and accounts
receivable, the accounting department determines that
collection experience is as shown in Table 4.18. For
example, 75% of the IBM adapters sold in a quarter
are paid for during the quarter, 20% are paid for dur-
ing the following quarter, and 3% are paid for during
212 Chapter 4 Linear Programming Models
the third quarter. The remaining 2% are written off
and sold to a collection agency for $0.50 on the dollar.
Table 4.18 Collections
Quarter IBM Adapters Macintosh Adapters
1 0.75 0.80
2 0.20 0.11
3 0.03 0.05
Questions
1. Suppose that you are Garcia. Develop an LP
spreadsheet model of the situation to help the
executive action team make a decision about
how to allocate funds between production and
advertising so that all the cash flow, financial,
accounting, marketing, inventory, and production
considerations are taken into account and
American Office Systems’ profits are maximized.
Use the data collected and the estimates made
by the members of the executive action team.
2. Solve the LP model formulated in Question 1.
The executive action team has assembled to reconsider
the plans for the adapters for the coming year. Garcia, who
developed the LP model, concludes his presentation by
saying,“As everyone can see, the model gives the optimal
solution that maximizes profits. Since I have incorporated
the estimates and assumptions that all of you made,
clearly it is the best solution. No other alternative can give
a higher profit.” Even Williams, who initially was skeptical
of using quantitative models for making executive-level
decisions, is impressed and indicates that she will go along
with the results.
Miller says,“Good work, Carlos! This is a complex
problem but your presentation made it all seem so simple.
However, remember that those figures you used were
based on estimates made by all of us. Some were little bet-
ter than guesses. What happens if they are wrong? In
other words, your presentation has helped me get a handle
on the problem we are facing, and I know that models are
useful where hard, accurate, data exist. However, with all
the uncertainty in our situation and the many rough esti-
mates made, it seems to me that I will still have to make a
judgment call when it comes down to making a final
decision. Also, there has been a new development. J.T. tells
me that we might be able to get another $1 million
line of credit from a Bahamian bank. It will take a while
to work out the details and maybe it will cost us a little.
I am wondering if it is worth it. What would we do with
the $1 million if we got it?” T. J. responds,“We really need
the $1 million. But it is a drop in the bucket. My analysis
shows that we really need another $8 million line of
credit.”
Analyze, as Garcia is going to do, the effect of uncer-
tainty and errors on the results of Questions 1 and 2 by
answering the following questions.They should be
attempted only after Questions 1 and 2 have been
answered correctly.
3. One area where assumptions were made is
adapter price.
a. What happens if the prices for the adapters
are a little weak and they decrease to $173
for the IBM version and $198 for the
Macintosh version? Does this make any
difference?
b. What about decreases to $172 and $197,
respectively, for the IBM and Macintosh
versions? Explain the answers in terms that
Miller will understand.
c. Suppose that American Office Systems can
increase the price of the adapters to $180
and $205. How would this affect the original
solution?
4. Another potential variable is adapter production
cost.
a. Suppose that an error was made in determin-
ing the costs of the adapters and that they
really should have been $102 for the IBM
version and $112 for the Macintosh version.
What is the effect of this error?
b. What about costs of $105 and $115? Explain
the answers in terms that Miller will
understand.
5. Howell notes that one of the contributing factors
to American Office Systems’ cash squeeze is the
slow collection of accounts receivable. He is con-
sidering adopting a new collection procedure
recommended by a consulting company. It will
cost $100,000 and will change the collection
rates to those given in Table 4.19.
a. Analyze the effect of this new collection
policy and make a recommendation to Howell
about whether to implement the new proce-
dure. As before, any accounts receivable not
collected by the end of the third quarter will
be sold to a collection agency for $0.50 on
the dollar.
b. Howell wonders whether switching to selling
adapters for all cash is worth the effort. This
would ameliorate the cash squeeze because it
would eliminate not only the slow collections
but also the use of the collection agency for
accounts that remain unpaid after 9 months.
It would cost about $90,000 more than at
present to implement the all-cash policy be-
cause the accounting system would need to
be modified and personnel would have to be
retrained. Analyze this possibility and make a
recommendation to Howell.
Table 4.19 New Collections
Quarter IBM Adapters Macintosh Adapters
1 0.90 0.92
2 0.07 0.03
3 0.01 0.01
6. Yet another variable is advertising effectiveness.
a. Suppose that Williams overestimated the
effectiveness of advertising. It now appears
that $100 is needed to increase sales by one
adapter. How will this affect the original solu-
tion? Explain the answer in terms that Miller
will understand.
b. What happens if the required advertising
outlay is $12.50 per additional adapter sold?
7. Suppose that the line of credit from Citibank that
Howell thought he had arranged did not work
out because of the poor financial situation of the
company. The company can obtain one for the
same amount from a small local bank; however,
the interest rate is much higher, 24%. Analyze
how this change affects American Office Systems.
Case 4.2 American Office Systems, Inc. 213
8. The safety cushion for inventory is subject to
revision.
a. Suppose that Garcia finds a bug in his original
inventory model. Correcting it results in a
safety cushion of 15% instead of the 10% he
suggested previously. Determine whether this
is important.
b. What if the error is 20%? Explain the answers
in terms that Miller will understand.
9. Production capacity is scheduled to increase by
10% in the fourth quarter.
a. Suppose that Miller is advised by the
construction company that the work will not
be finished until the following year. How will
this delay affect the company’s plans?
b. In addition to the delay in part a, suppose
that an accident in the production facility
damages some of the equipment so that the
capacity is decreased by 10% in the fourth
quarter. Analyze how this will affect the
original solution.
214 Chapter 4 Linear Programming Models
10. Williams is worried about the accuracy of
Lu’s 1996 maximum expected sales forecasts. If
errors in these forecasts have a big effect on the
company profits, she is thinking about hiring a
San Francisco marketing research firm to do a
more detailed analysis. They would charge
$50,000 for a study. Help Williams by analyzing
what would happen if Lu’s forecasts are in error
by 1,000 for IBM adapters and 200 for Macin-
tosh adapters each quarter. Should she hire the
marketing research firm?
11. a. To determine whether the extra $1 million
line of credit is needed, analyze its effect on
the original solution given in Question 2.
b. To fully understand the ramifications of the
extra $1,000,000 line of credit, redo (1) Ques-
tion 3b, (2) Question 4b, (3) Question 6a, and
(4) Question 8b. Summarize your results.
c. What about Howell’s claim that an extra
$8,000,000 line of credit is necessary? Use
that adjustment and redo Question 6a. ■
3/17/2020 Case_Study_Worksheet - final 1.0

MGSC 404 Foundations of Decision Analytics

Case Study Worksheet

Team Members: Ammar Khan, Alexandre Chahtahtinsky, William Horwitz, Maguette Paye, Matanel
Cohen-Weinberg, Maxence Prémont

In [5]: from gurobipy import *

from datascience import *
import numpy as np
%matplotlib inline
import matplotlib.pyplot as plots
plots.style.use('fivethirtyeight')
import math
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

Question 1
Assumptions

Our model relies on the following assumptions:

1. American Office Systems, Inc. runs 2 work shifts of 8 hours every day.
2. The IBM and Macintosh adapters sell for $175 and $200 respectively
3. The IBM and Macintosh adapters cost $100 and $110 respectively

Decision Variables

We define the following decision variables:

1. P
A
i
= IBM adapters [in units] produced in quarter i, i = 1, 2, 3, 4
2. P
B
i
= Macintosh adaptors [in units] produced in quarter i, i = 1, 2, 3, 4
3. M
i
A
= IBM adapters marketing expense [in $] in quarter i, i = 1, 2, 3, 4
4. M
i
B
= Macintosh adapters marketing expense [in $] in quarter i, i = 1, 2, 3, 4

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 1/28

3/17/2020 Case_Study_Worksheet - final 1.0

Parameters

We define the following parameters:

1. Q
A
i
= IBM adapters [in units] sold in quarter i, i = 1, 2, 3, 4
2. Q
B
i
= Macintosh adapters [in units] sold in quarter i, i = 1, 2, 3, 4
3. I
A
i
= IBM adapters inventory [in units] at the end of quarter i, i = 1, 2, 3, 4
4. I
B
i
= Macintosh adapters inventory [in units] at the end of quarter i, i = 1, 2, 3, 4

Objective Function

Maximize Total Profits

A B
max (200 ∗ Q ) + (175 ∗ Q ) − Costs
i i
A B ∑ ( )
Q ,Q
i∈{1,2,3,4}

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 2/28

3/17/2020 Case_Study_Worksheet - final 1.0

Constraints
Production Constraints:

1. Pi
A
≤ 28800 maximum IBM adapters [in units] produced in quarter i, i = 1, 2, 3
2. P
A
4
≤ 31680 maximum IBM adapters [in units] produced in the fourth quarter

3. P
B
i
≤ 9600 maximum Macintosh adapters [in units] produced in quarter i, i = 1, 2, 3

4. P
B
4
≤ 10560 maximum Macintosh adapters [in units] produced in the fourth quarter

Invenotry Constraints

1. P
A
i
+ I iA ≥ (QA
j
)*1.1 necessary IBM adapters production for quarter i, i = 1, 2, 3 , where j = i + 1

2. P
A
4
+ I4A
≥ (Q4 )*2 necessary IBM adapters production for fourth quarter
A

3. P
B
i
+ I iB ≥ (Q )*1.1 necessary Macintosh adapters production for quarter i, i
B
j
= 1, 2, 3 , where
j = i + 1

4. P
B
4
+ I 4B ≥ (QB4
)*2 necessary Macintosh adapters production for fourth quarter
5. I
A
i
= I k + Pi - QA
A A
i
ending inventory for IBM adapters [in units] for quarter i, i = 2, 3, 4 , where
k = i − 1

6. I
A
1
= 10000 + PA 1
- QA
1
ending inventory for IBM adapters [in units] for the first quarter
7. I
B
i
= I k + Pi - Qi ending inventory for Macintosh adapters [in units] for quarter i, i =
B B B
2, 3, 4 , where
k = i − 1

8. I
B
1
= 5000 + PB
1
- QB
1
ending inventory for Macintosh adapters [in units] for the first quarter

Quantity Sold Constraints:

1. Q
A
i
= NiA + MiA /10 IBM adapter sold [in units] for quarter i, i = 1, 2, 3, 4

2. Q
B
i
= NiB + MiB /10 IBM adapter sold [in units] for quarter i, i = 1, 2, 3, 4

Budget Constraints:

1. P
A
i
*100 + MiA + PB
i
*110 + MiB ≤ 750000 for quarter i, i = 1, 2, 3, 4

In [6]: prod_capacity_a = [28800,28800,28800,31680]

prod_capacity_b = [9600,9600,9600,10560]
base_sales_a = [9000,10000,11000,12000]
base_sales_b = [1800,2000,2200,2400]

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 3/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [7]: model = Model('Case Study V1')

P_a={} ##Production for IBM

P_b={} ##Production for Macintosh
Q_a={} ##Quantity sold for IBM
Q_b={} ##Quantity sold for Mac
I_a={} ##End of period inventory for IBM
I_b={} ##End of period inventory for Macintosh
M_a={} ##Marketing expense for IBM
M_b={} ##Marketing expense for Macintosh
B={} ##Loan
C={} ##Cash on hand at quarter end
X={} ##Excess Cash

##Creating the production,Quantity sold, and end of period inventory variables

for i in np.arange(4):
P_a[i] = model.addVar(lb=0)
P_b[i] = model.addVar(lb=0)
Q_a[i] = model.addVar(lb=0)
Q_b[i] = model.addVar(lb=0)
I_a[i] = model.addVar(lb=0)
I_b[i] = model.addVar(lb=0)
M_a[i] = model.addVar(lb=0)
M_b[i] = model.addVar(lb=0)
B[i] = model.addVar(lb=0)
C[i] = model.addVar(lb=0)
X[i] = model.addVar(lb=0)

##Creating the constraints for production

con_production ={}
##For IBM
for i in np.arange(4):
con_production[i]=model.addConstr(P_a[i]<=prod_capacity_a[i])

##For Macintosh
for i in np.arange(4):
con_production[i+4]=model.addConstr(P_b[i]<=prod_capacity_b[i])

##Creating the constraints for end of period inventory

con_inventory={}
for i in np.arange(4):
if i==0:
con_inventory[i]=model.addConstr(I_a[i]==10000+P_a[i]-Q_a[i])##Quarter
1 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==5000+P_b[i]-Q_b[i])##Quarte
r 1 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
elif i<3:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##Quat
er 2-3 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])##Qu
arter 2-3 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
else:

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 4/28

3/17/2020 Case_Study_Worksheet - final 1.0

con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##Quat
er 4 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])##Qu
arter 4 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=2*Q_a[i])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=2*Q_b[i])

##Creating the constraints for Quantity sold

con_sales={}
##For IBM
for i in np.arange(4):
con_sales[i]=model.addConstr(Q_a[i]==base_sales_a[i]+M_a[i]*0.1) ##For IBM
con_sales[i+4]=model.addConstr(Q_b[i]==base_sales_b[i]+M_b[i]*0.1) ## For
Macintosh

##Creating the constraints for the loan

con_loan={}
for i in np.arange(4):
con_loan[i]=model.addConstr(B[i]<=750000)

##Creating constraints for the cashflow

con_cashflow={}
con_cashflow[0]=model.addConstr(C[0]==B[0]+(175*0.75*Q_a[0] + 175*0.2*8000 + 1
75*0.03*7000 + 175*0.02*0.5*7000)
+(200*0.8*Q_b[0] + 200*0.11*1600 + 200*0.05*140
0 + 200*0.04*0.5*1400)
-(100*P_a[0])-(110*P_b[0])-(100*0.03*I_a[0])-(1
10*0.03*I_b[0])
-M_a[0] - M_b[0]-500000)
con_cashflow[1]=model.addConstr(C[1]==B[1]+(175*0.75*Q_a[1] + 175*0.2*Q_a[0] +
175*0.03*8000 + 175*0.02*0.5*8000)
+(200*0.8*Q_b[1] + 200*0.11*Q_b[0] + 200*0.05*1
600 + 200*0.04*0.5*1600)
-(100*P_a[1])-(110*P_b[1])-(100*0.03*I_a[1])-(1
10*0.03*I_b[1])
-M_a[1] - M_b[1] - B[0]*1.04-500000)
con_cashflow[2]=model.addConstr(C[2]==B[2]+(175*0.75*Q_a[2] + 175*0.2*Q_a[1] +
175*0.03*Q_a[0] + 175*0.02*0.5*Q_b[0])
+(200*0.8*Q_b[2] + 200*0.11*Q_b[1] + 200*0.05*Q
_b[0] + 200*0.04*0.5*Q_b[0])
-(100*P_a[2])-(110*P_b[2])-(100*0.03*I_a[2])-(1
10*0.03*I_b[2])
-M_a[2] - M_b[2] - B[1]*1.04-500000)
con_cashflow[3]=model.addConstr(C[3]==B[3]+(175*0.75*Q_a[3] + 175*0.2*Q_a[2] +
175*0.03*Q_a[1] + 175*0.02*0.5*Q_b[1])
+(200*0.8*Q_b[3] + 200*0.11*Q_b[2] + 200*0.05*Q
_b[1] + 200*0.04*0.5*Q_b[1])
-(100*P_a[3])-(110*P_b[3])-(100*0.03*I_a[3])-(1
10*0.03*I_b[3])
-M_a[3] - M_b[3] - B[2]*1.04-500000)

##Creating constraints for the excess cash

con_excess_cash={}
for i in np.arange(4):
if C[i]>=0:
con_excess_cash[i]=model.addConstr(X[i]==C[i])
else: con_excess_cash[i]=model.addConstr(X[i]==0)

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 5/28

3/17/2020 Case_Study_Worksheet - final 1.0

con_negative={}
for i in np.arange(4):
con_negative[i]=model.addConstr(P_a[i]>=0)
con_negative[i+4]=model.addConstr(P_b[i]>=0)
con_negative[i+8]=model.addConstr(M_a[i]>=0)
con_negative[i+12]=model.addConstr(M_b[i]>=0)
con_negative[i+16]=model.addConstr(B[i]>=0)

Academic license - for non-commercial use only

In [8]: ##Objective Function

obj = C[0]+C[1]+C[2]+C[3]+(X[0]*0.045+X[1]*0.03+X[2]*0.015)-(B[3]*1.04/1.06**(
1/4))+(175*0.2*Q_a[3] + 175*0.03*Q_a[2] + 175*0.02*0.5*Q_b[2])/(1.06**(1/4))+(
175*0.03*Q_a[3] + 175*0.02*0.5*Q_b[3])/(1.06**(2/4))+ (200*0.11*Q_b[3] + 200*
0.05*Q_b[2] + 200*0.04*0.5*Q_b[2])/(1.06**(1/4))+ (200*0.05*Q_b[3] + 200*0.04*
0.5*Q_b[3])/(1.06**(2/4))
model.setObjective(obj, GRB.MAXIMIZE)

model.optimize()

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [1e-01, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.01s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.8976000e+32 3.000000e+30 3.897600e+02 0s
16 9.3764381e+06 0.000000e+00 0.000000e+00 0s

Solved in 16 iterations and 0.02 seconds

Optimal objective 9.376438110e+06

In [9]: ##Results
inv_table_a=Table().with_columns('Quarters',[1,2,3,4],'Beginnig Inventory',[10
000,I_a[0].x,I_a[1].x,I_a[2].x]
,'Production IBM',[P_a[0].x,P_a[1].x,P_a[2].x,P_a[3].x]
,'Quantity Sold',[Q_a[0].x,Q_a[1].x,Q_a[2].x,Q_a[3].x]
,'Ending Inventory',[I_a[0].x,I_a[1].x,I_a[2].x,I_a[3].x])
inv_table_a

Out[9]: Quarters Beginnig Inventory Production IBM Quantity Sold Ending Inventory

1 10000 28800 38800 0

2 0 28800 26181.8 2618.18

3 2618.18 25943.8 28562 0

4 0 31680 21120 10560

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 6/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [10]: inv_table_b=Table().with_columns('Quarters',[1,2,3,4],'Beginnig Inventory',[50

00,I_b[0].x,I_b[1].x,I_b[2].x]
,'Production MAC',[P_b[0].x,P_b[1].x,P_b[2].x,P_b[3].x]
,'Quantity Sold',[Q_b[0].x,Q_b[1].x,Q_b[2].x,Q_b[3].x]
,'Ending Inventory',[I_b[0].x,I_b[1].x,I_b[2].x,I_b[3].x])
inv_table_b

Out[10]: Quarters Beginnig Inventory Production MAC Quantity Sold Ending Inventory

1 5000 9600 14600 0

2 0 9600 8727.27 872.727

3 872.727 8647.93 9520.66 0

4 0 10560 7040 3520

Question 2
In [11]: Solution = Table().with_columns([
'Quarters',[1,2,3,4],
'IBM Production',[P_a[0].x,P_a[1].x,P_a[2].x,P_a[3].x],
'Macintosh Production',[P_b[0].x,P_b[1].x,P_b[2].x,P_b[3].x],
'Marketing Expense IBM',[M_a[0].x,M_a[1].x,M_a[2].x,M_a[3].x],
'Marketing Expense Mac',[M_b[0].x,M_b[1].x,M_b[2].x,M_b[3].x]
])
Solution ## Multiply productions by 100 and 110.

Out[11]: IBM Macintosh Marketing Expense Marketing Expense

Quarters
Production Production IBM Mac

1 28800 9600 298000 128000

2 28800 9600 161818 67272.7

3 25943.8 8647.93 175620 73206.6

4 31680 10560 91200 46400

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 7/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [12]: plt.plot( 'Quarters', 'IBM Production', data=Solution, marker='o', markerfacec

olor='blue', markersize=12, color='skyblue', linewidth=4,label="IBM")
plt.plot( 'Quarters', 'Macintosh Production', data=Solution, marker='o', color
='red', markersize=12,linewidth=4, label="Macintosh")
plt.legend() ## Production Graph - label

Out[12]: <matplotlib.legend.Legend at 0x25011cc9d30>

In [13]: plt.plot( 'Quarters', 'Marketing Expense IBM', data=Solution, marker='o', mark

erfacecolor='blue', markersize=12, color='skyblue', linewidth=4,label="IBM")
plt.plot( 'Quarters', 'Marketing Expense Mac', data=Solution, marker='o', colo
r='red', markersize=12,linewidth=4, label="Macintosh")
plt.legend() ## Marketing graph - label

Out[13]: <matplotlib.legend.Legend at 0x25011fb22e8>

Other Questions
In [ ]: ##We need to answer Questions 3,4,6,7,10 and 2 original sensitivity analysis

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 8/28

3/17/2020 Case_Study_Worksheet - final 1.0

Question 3
Selling Price Sensitivity

What if the price of adapters were different?

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 9/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [14]: def sensitivity_price(price_IBM,price_MAC):

model=Model('Price Senstivity Model')

P_a={} ##Production for IBM

P_b={} ##Production for Macintosh
Q_a={} ##Quantity sold for IBM
Q_b={} ##Quantity sold for Mac
I_a={} ##End of period inventory for IBM
I_b={} ##End of period inventory for Macintosh
M_a={} ##Marketing expense for IBM
M_b={} ##Marketing expense for Macintosh
B={} ##Loan
C={} ##Cash on hand at quarter end
X={} ##Excess Cash

##Creating the production,Quantity sold, and end of period inventory varia

bles
for i in np.arange(4):
P_a[i] = model.addVar(lb=0)
P_b[i] = model.addVar(lb=0)
Q_a[i] = model.addVar(lb=0)
Q_b[i] = model.addVar(lb=0)
I_a[i] = model.addVar(lb=0)
I_b[i] = model.addVar(lb=0)
M_a[i] = model.addVar(lb=0)
M_b[i] = model.addVar(lb=0)
B[i] = model.addVar(lb=0)
C[i] = model.addVar(lb=0)
X[i] = model.addVar(lb=0)

##Creating the constraints for production

con_production ={}
##For IBM
for i in np.arange(4):
con_production[i]=model.addConstr(P_a[i]<=prod_capacity_a[i])

##For Macintosh
for i in np.arange(4):
con_production[i+4]=model.addConstr(P_b[i]<=prod_capacity_b[i])

##Creating the constraints for end of period inventory

con_inventory={}
for i in np.arange(4):
if i==0:
con_inventory[i]=model.addConstr(I_a[i]==10000+P_a[i]-Q_a[i])##Qua
rter 1 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==5000+P_b[i]-Q_b[i])##Qu
arter 1 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
elif i<3:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 2-3 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 2-3 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 10/28

3/17/2020 Case_Study_Worksheet - final 1.0

con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
else:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 4 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 4 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=2*Q_a[i])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=2*Q_b[i])

##Creating the constraints for Quantity sold

con_sales={}
##For IBM
for i in np.arange(4):
con_sales[i]=model.addConstr(Q_a[i]==base_sales_a[i]+M_a[i]*0.1) ##For
IBM
con_sales[i+4]=model.addConstr(Q_b[i]==base_sales_b[i]+M_b[i]*0.1) ##
For Macintosh

##Creating the constraints for the loan

con_loan={}
for i in np.arange(4):
con_loan[i]=model.addConstr(B[i]<=750000)

##Creating constraints for the cashflow

con_cashflow={}
con_cashflow[0]=model.addConstr(C[0]==B[0]+(price_IBM*0.75*Q_a[0] + price_
IBM*0.2*8000 + price_IBM*0.03*7000 + price_IBM*0.02*0.5*7000)
+(price_MAC*0.8*Q_b[0] + price_MAC*0.11*160
0 + price_MAC*0.05*1400 + price_MAC*0.04*0.5*1400)
-(100*P_a[0])-(110*P_b[0])-(100*0.03*I_a[0
])-(110*0.03*I_b[0])
-M_a[0] - M_b[0]-500000)
con_cashflow[1]=model.addConstr(C[1]==B[1]+(price_IBM*0.75*Q_a[1] + price_
IBM*0.2*Q_a[0] + price_IBM*0.03*8000 + price_IBM*0.02*0.5*8000)
+(price_MAC*0.8*Q_b[1] + price_MAC*0.11*Q_b
[0] + price_MAC*0.05*1600 + price_MAC*0.04*0.5*1600)
-(100*P_a[1])-(110*P_b[1])-(100*0.03*I_a[1
])-(110*0.03*I_b[1])
-M_a[1] - M_b[1] - B[0]*1.04-500000)
con_cashflow[2]=model.addConstr(C[2]==B[2]+(price_IBM*0.75*Q_a[2] + price_
IBM*0.2*Q_a[1] + price_IBM*0.03*Q_a[0] + price_IBM*0.02*0.5*Q_b[0])
+(price_MAC*0.8*Q_b[2] + price_MAC*0.11*Q_b
[1] + price_MAC*0.05*Q_b[0] + price_MAC*0.04*0.5*Q_b[0])
-(100*P_a[2])-(110*P_b[2])-(100*0.03*I_a[2
])-(110*0.03*I_b[2])
-M_a[2] - M_b[2] - B[1]*1.04-500000)
con_cashflow[3]=model.addConstr(C[3]==B[3]+(price_IBM*0.75*Q_a[3] + price_
IBM*0.2*Q_a[2] + price_IBM*0.03*Q_a[1] + price_IBM*0.02*0.5*Q_b[1])
+(price_MAC*0.8*Q_b[3] + price_MAC*0.11*Q_b
[2] + price_MAC*0.05*Q_b[1] + price_MAC*0.04*0.5*Q_b[1])
-(100*P_a[3])-(110*P_b[3])-(100*0.03*I_a[3
])-(110*0.03*I_b[3])
-M_a[3] - M_b[3] - B[2]*1.04-500000)

##Creating constraints for the excess cash

con_excess_cash={}
for i in np.arange(4):

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 11/28

3/17/2020 Case_Study_Worksheet - final 1.0

if C[i]>=0:
con_excess_cash[i]=model.addConstr(X[i]==C[i])
else: con_excess_cash[i]=model.addConstr(X[i]==0)

con_negative={}
for i in np.arange(4):
con_negative[i]=model.addConstr(P_a[i]>=0)
con_negative[i+4]=model.addConstr(P_b[i]>=0)
con_negative[i+8]=model.addConstr(M_a[i]>=0)
con_negative[i+12]=model.addConstr(M_b[i]>=0)
con_negative[i+16]=model.addConstr(B[i]>=0)

obj = C[0]+C[1]+C[2]+C[3]+(X[0]*0.045+X[1]*0.03+X[2]*0.015)-(B[3]*1.04/1.0
6**(1/4))+(price_IBM*0.2*Q_a[3] + price_IBM*0.03*Q_a[2] + price_IBM*0.02*0.5*Q
_b[2])/(1.06**(1/4))+(price_IBM*0.03*Q_a[3] + price_IBM*0.02*0.5*Q_b[3])/(1.06
**(2/4))+ (price_MAC*0.11*Q_b[3] + price_MAC*0.05*Q_b[2] + price_MAC*0.04*0.5*
Q_b[2])/(1.06**(1/4))+ (200*0.05*Q_b[3] + 200*0.04*0.5*Q_b[3])/(1.06**(2/4))
model.setObjective(obj, GRB.MAXIMIZE)

model.optimize()

return model.objVal

The price for adapters decreases to $173 for IBM and to $198 for Macintosh

In [15]: profit_sensitivity_1=sensitivity_price(173,198)
profit_sensitivity_1;

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [1e-01, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.01s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.8976000e+32 3.000000e+30 3.897600e+02 0s
16 9.0615883e+06 0.000000e+00 0.000000e+00 0s

Solved in 16 iterations and 0.02 seconds

Optimal objective 9.061588340e+06

In [16]: sen_p_1 = [173,198,profit_sensitivity_1]

sen_p_1

Out[16]: [173, 198, 9061588.340067077]

In [17]: x=[175,173,172,180]
y=[200,198,197,205]
z=[950,930,925,975]

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 12/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [279]: ax = plt.axes(projection='3d')

ax.scatter3D(x, y, z, c="red");

The price for adapters decreases to $172 for IBM and to $197 for Macintosh

In [239]: sensitivity_price(172,197)

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [1e-01, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.01s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.8976000e+32 3.000000e+30 3.897600e+02 0s
16 8.9041635e+06 0.000000e+00 0.000000e+00 0s

Solved in 16 iterations and 0.02 seconds

Optimal objective 8.904163455e+06
Out[239]: 8904163.454891495

The price for adapters increasesto $180 for IBM and to $205 for Macintosh

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 13/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [240]: sensitivity_price(180,205)

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [1e-01, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.01s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.8976000e+32 3.000000e+30 3.897600e+02 0s
16 1.0163563e+07 0.000000e+00 0.000000e+00 0s

Solved in 16 iterations and 0.01 seconds

Optimal objective 1.016356254e+07
Out[240]: 10163562.536296174

Question 4
Prodcution Cost Sensitivity

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 14/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [219]: def sensitivity_vc(vc_IBM,vc_MAC):

model=Model('Production Cost Senstivity Model')

P_a={} ##Production for IBM

P_b={} ##Production for Macintosh
Q_a={} ##Quantity sold for IBM
Q_b={} ##Quantity sold for Mac
I_a={} ##End of period inventory for IBM
I_b={} ##End of period inventory for Macintosh
M_a={} ##Marketing expense for IBM
M_b={} ##Marketing expense for Macintosh
B={} ##Loan
C={} ##Cash on hand at quarter end
X={} ##Excess Cash

##Creating the production,Quantity sold, and end of period inventory varia

bles
for i in np.arange(4):
P_a[i] = model.addVar(lb=0)
P_b[i] = model.addVar(lb=0)
Q_a[i] = model.addVar(lb=0)
Q_b[i] = model.addVar(lb=0)
I_a[i] = model.addVar(lb=0)
I_b[i] = model.addVar(lb=0)
M_a[i] = model.addVar(lb=0)
M_b[i] = model.addVar(lb=0)
B[i] = model.addVar(lb=0)
C[i] = model.addVar(lb=0)
X[i] = model.addVar(lb=0)

##Creating the constraints for production

con_production ={}
##For IBM
for i in np.arange(4):
con_production[i]=model.addConstr(P_a[i]<=prod_capacity_a[i])

##For Macintosh
for i in np.arange(4):
con_production[i+4]=model.addConstr(P_b[i]<=prod_capacity_b[i])

##Creating the constraints for end of period inventory

con_inventory={}
for i in np.arange(4):
if i==0:
con_inventory[i]=model.addConstr(I_a[i]==10000+P_a[i]-Q_a[i])##Qua
rter 1 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==5000+P_b[i]-Q_b[i])##Qu
arter 1 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
elif i<3:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 2-3 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 2-3 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 15/28

3/17/2020 Case_Study_Worksheet - final 1.0

con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
else:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 4 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 4 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=2*Q_a[i])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=2*Q_b[i])

##Creating the constraints for Quantity sold

con_sales={}
##For IBM
for i in np.arange(4):
con_sales[i]=model.addConstr(Q_a[i]==base_sales_a[i]+M_a[i]*0.1) ##For
IBM
con_sales[i+4]=model.addConstr(Q_b[i]==base_sales_b[i]+M_b[i]*0.1) ##
For Macintosh

##Creating the constraints for the loan

con_loan={}
for i in np.arange(4):
con_loan[i]=model.addConstr(B[i]<=750000)

##Creating constraints for the cashflow

con_cashflow={}
con_cashflow[0]=model.addConstr(C[0]==B[0]+(175*0.75*Q_a[0] + 175*0.2*8000
+ 175*0.03*7000 + 175*0.02*0.5*7000)
+(200*0.8*Q_b[0] + 200*0.11*1600 + 200*0.05*140
0 + 200*0.04*0.5*1400)
-(vc_IBM*P_a[0])-(vc_MAC*P_b[0])-(vc_IBM*0.03*I
_a[0])-(vc_MAC*0.03*I_b[0])
-M_a[0] - M_b[0]-500000)
con_cashflow[1]=model.addConstr(C[1]==B[1]+(175*0.75*Q_a[1] + 175*0.2*Q_a[
0] + 175*0.03*8000 + 175*0.02*0.5*8000)
+(200*0.8*Q_b[1] + 200*0.11*Q_b[0] + 200*0.05*1
600 + 200*0.04*0.5*1600)
-(vc_IBM*P_a[1])-(vc_MAC*P_b[1])-(vc_IBM*0.03*I
_a[1])-(vc_MAC*0.03*I_b[1])
-M_a[1] - M_b[1] - B[0]*1.04-500000)
con_cashflow[2]=model.addConstr(C[2]==B[2]+(175*0.75*Q_a[2] + 175*0.2*Q_a[
1] + 175*0.03*Q_a[0] + 175*0.02*0.5*Q_b[0])
+(200*0.8*Q_b[2] + 200*0.11*Q_b[1] + 200*0.05*Q
_b[0] + 200*0.04*0.5*Q_b[0])
-(vc_IBM*P_a[2])-(vc_MAC*P_b[2])-(vc_IBM*0.03*I
_a[2])-(vc_MAC*0.03*I_b[2])
-M_a[2] - M_b[2] - B[1]*1.04-500000)
con_cashflow[3]=model.addConstr(C[3]==B[3]+(175*0.75*Q_a[3] + 175*0.2*Q_a[
2] + 175*0.03*Q_a[1] + 175*0.02*0.5*Q_b[1])
+(200*0.8*Q_b[3] + 200*0.11*Q_b[2] + 200*0.05*Q
_b[1] + 200*0.04*0.5*Q_b[1])
-(vc_IBM*P_a[3])-(vc_MAC*P_b[3])-(vc_IBM*0.03*I
_a[3])-(vc_MAC*0.03*I_b[3])
-M_a[3] - M_b[3] - B[2]*1.04-500000)

##Creating constraints for the excess cash

con_excess_cash={}
for i in np.arange(4):

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 16/28

3/17/2020 Case_Study_Worksheet - final 1.0

if C[i]>=0:
con_excess_cash[i]=model.addConstr(X[i]==C[i])
else: con_excess_cash[i]=model.addConstr(X[i]==0)

con_negative={}
for i in np.arange(4):
con_negative[i]=model.addConstr(P_a[i]>=0)
con_negative[i+4]=model.addConstr(P_b[i]>=0)
con_negative[i+8]=model.addConstr(M_a[i]>=0)
con_negative[i+12]=model.addConstr(M_b[i]>=0)
con_negative[i+16]=model.addConstr(B[i]>=0)

obj = C[0]+C[1]+C[2]+C[3]+(X[0]*0.045+X[1]*0.03+X[2]*0.015)-(B[3]*1.04/1.0
6**(1/4))+(175*0.2*Q_a[3] + 175*0.03*Q_a[2] + 175*0.02*0.5*Q_b[2])/(1.06**(1/4
))+(175*0.03*Q_a[3] + 175*0.02*0.5*Q_b[3])/(1.06**(2/4))+ (200*0.11*Q_b[3] + 2
00*0.05*Q_b[2] + 200*0.04*0.5*Q_b[2])/(1.06**(1/4))+ (200*0.05*Q_b[3] + 200*0.
04*0.5*Q_b[3])/(1.06**(2/4))
model.setObjective(obj, GRB.MAXIMIZE)

model.optimize()

return model.objVal

The production cost increase to $102 for IBM and $112 for Macintosh

In [220]: sensitivity_vc(102,112)

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [1e-01, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.01s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.8976000e+32 3.000000e+30 3.897600e+02 0s
16 9.0613163e+06 0.000000e+00 0.000000e+00 0s

Solved in 16 iterations and 0.02 seconds

Optimal objective 9.061316349e+06
Out[220]: 9061316.349095933

The production cost increases to $105 for IBM and $115 for Maintosh

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 17/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [221]: sensitivity_vc(105,115)

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [1e-01, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.01s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.8976000e+32 3.000000e+30 3.897600e+02 0s
16 8.5886337e+06 0.000000e+00 0.000000e+00 0s

Solved in 16 iterations and 0.02 seconds

Optimal objective 8.588633707e+06
Out[221]: 8588633.707112465

Question 6
Advertising Effectiveness Sensitivity

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 18/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [225]: def sensitivity_marketing(advert_effect):

model=Model('Marketing Senstivity Model')

P_a={} ##Production for IBM

P_b={} ##Production for Macintosh
Q_a={} ##Quantity sold for IBM
Q_b={} ##Quantity sold for Mac
I_a={} ##End of period inventory for IBM
I_b={} ##End of period inventory for Macintosh
M_a={} ##Marketing expense for IBM
M_b={} ##Marketing expense for Macintosh
B={} ##Loan
C={} ##Cash on hand at quarter end
X={} ##Excess Cash

##Creating the production,Quantity sold, and end of period inventory varia

bles
for i in np.arange(4):
P_a[i] = model.addVar(lb=0)
P_b[i] = model.addVar(lb=0)
Q_a[i] = model.addVar(lb=0)
Q_b[i] = model.addVar(lb=0)
I_a[i] = model.addVar(lb=0)
I_b[i] = model.addVar(lb=0)
M_a[i] = model.addVar(lb=0)
M_b[i] = model.addVar(lb=0)
B[i] = model.addVar(lb=0)
C[i] = model.addVar(lb=0)
X[i] = model.addVar(lb=0)

##Creating the constraints for production

con_production ={}
##For IBM
for i in np.arange(4):
con_production[i]=model.addConstr(P_a[i]<=prod_capacity_a[i])

##For Macintosh
for i in np.arange(4):
con_production[i+4]=model.addConstr(P_b[i]<=prod_capacity_b[i])

##Creating the constraints for end of period inventory

con_inventory={}
for i in np.arange(4):
if i==0:
con_inventory[i]=model.addConstr(I_a[i]==10000+P_a[i]-Q_a[i])##Qua
rter 1 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==5000+P_b[i]-Q_b[i])##Qu
arter 1 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
elif i<3:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 2-3 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 2-3 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 19/28

3/17/2020 Case_Study_Worksheet - final 1.0

con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
else:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 4 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 4 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=2*Q_a[i])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=2*Q_b[i])

##Creating the constraints for Quantity sold

con_sales={}
##For IBM
for i in np.arange(4):
con_sales[i]=model.addConstr(Q_a[i]==base_sales_a[i]+M_a[i]/advert_eff
ect) ##For IBM
con_sales[i+4]=model.addConstr(Q_b[i]==base_sales_b[i]+M_b[i]/advert_e
ffect) ## For Macintosh

##Creating the constraints for the loan

con_loan={}
for i in np.arange(4):
con_loan[i]=model.addConstr(B[i]<=750000)

##Creating constraints for the cashflow

con_cashflow={}
con_cashflow={}
con_cashflow[0]=model.addConstr(C[0]==B[0]+(175*0.75*Q_a[0] + 175*0.2*8000
+ 175*0.03*7000 + 175*0.02*0.5*7000)
+(200*0.8*Q_b[0] + 200*0.11*1600 + 200*0.05*140
0 + 200*0.04*0.5*1400)
-(100*P_a[0])-(110*P_b[0])-(100*0.03*I_a[0])-(1
10*0.03*I_b[0])
-M_a[0] - M_b[0]-500000)
con_cashflow[1]=model.addConstr(C[1]==B[1]+(175*0.75*Q_a[1] + 175*0.2*Q_a[
0] + 175*0.03*8000 + 175*0.02*0.5*8000)
+(200*0.8*Q_b[1] + 200*0.11*Q_b[0] + 200*0.05*1
600 + 200*0.04*0.5*1600)
-(100*P_a[1])-(110*P_b[1])-(100*0.03*I_a[1])-(1
10*0.03*I_b[1])
-M_a[1] - M_b[1] - B[0]*1.04-500000)
con_cashflow[2]=model.addConstr(C[2]==B[2]+(175*0.75*Q_a[2] + 175*0.2*Q_a[
1] + 175*0.03*Q_a[0] + 175*0.02*0.5*Q_b[0])
+(200*0.8*Q_b[2] + 200*0.11*Q_b[1] + 200*0.05*Q
_b[0] + 200*0.04*0.5*Q_b[0])
-(100*P_a[2])-(110*P_b[2])-(100*0.03*I_a[2])-(1
10*0.03*I_b[2])
-M_a[2] - M_b[2] - B[1]*1.04-500000)
con_cashflow[3]=model.addConstr(C[3]==B[3]+(175*0.75*Q_a[3] + 175*0.2*Q_a[
2] + 175*0.03*Q_a[1] + 175*0.02*0.5*Q_b[1])
+(200*0.8*Q_b[3] + 200*0.11*Q_b[2] + 200*0.05*Q
_b[1] + 200*0.04*0.5*Q_b[1])
-(100*P_a[3])-(110*P_b[3])-(100*0.03*I_a[3])-(1
10*0.03*I_b[3])
-M_a[3] - M_b[3] - B[2]*1.04-500000)
##Creating constraints for the excess cash
con_excess_cash={}
for i in np.arange(4):

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 20/28

3/17/2020 Case_Study_Worksheet - final 1.0

if C[i]>=0:
con_excess_cash[i]=model.addConstr(X[i]==C[i])
else: con_excess_cash[i]=model.addConstr(X[i]==0)

con_negative={}
for i in np.arange(4):
con_negative[i]=model.addConstr(P_a[i]>=0)
con_negative[i+4]=model.addConstr(P_b[i]>=0)
con_negative[i+8]=model.addConstr(M_a[i]>=0)
con_negative[i+12]=model.addConstr(M_b[i]>=0)
con_negative[i+16]=model.addConstr(B[i]>=0)

obj = C[0]+C[1]+C[2]+C[3]+(X[0]*0.045+X[1]*0.03+X[2]*0.015)-(B[3]*1.04/1.0
6**(1/4))+(175*0.2*Q_a[3] + 175*0.03*Q_a[2] + 175*0.02*0.5*Q_b[2])/(1.06**(1/4
))+(175*0.03*Q_a[3] + 175*0.02*0.5*Q_b[3])/(1.06**(2/4))+ (200*0.11*Q_b[3] + 2
00*0.05*Q_b[2] + 200*0.04*0.5*Q_b[2])/(1.06**(1/4))+ (200*0.05*Q_b[3] + 200*0.
04*0.5*Q_b[3])/(1.06**(2/4))
model.setObjective(obj, GRB.MAXIMIZE)

model.optimize()

return model.objVal

The advertising effectiveness decreases to $100, where $100 can increase the demand by one unit

In [226]: sensitivity_marketing(100)

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [1e-02, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.01s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.1122074e+07 3.446418e+05 0.000000e+00 0s
19 2.7772152e+06 0.000000e+00 0.000000e+00 0s

Solved in 19 iterations and 0.02 seconds

Optimal objective 2.777215150e+06
Out[226]: 2777215.150349453

The advertising effectiveness decreases to $12.5, where $12.5 can increase the demand by one unit

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 21/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [228]: sensitivity_marketing(12.5)

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [8e-02, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.01s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.8976000e+32 3.000000e+30 3.897600e+02 0s
16 9.1086150e+06 0.000000e+00 0.000000e+00 0s

Solved in 16 iterations and 0.02 seconds

Optimal objective 9.108614991e+06
Out[228]: 9108614.990583539

Question 7
Credit Line Interest Sensitivity

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 22/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [231]: def sensitivity_interest(interest):

model=Model('Interest Senstivity Model')

P_a={} ##Production for IBM

P_b={} ##Production for Macintosh
Q_a={} ##Quantity sold for IBM
Q_b={} ##Quantity sold for Mac
I_a={} ##End of period inventory for IBM
I_b={} ##End of period inventory for Macintosh
M_a={} ##Marketing expense for IBM
M_b={} ##Marketing expense for Macintosh
B={} ##Loan
C={} ##Cash on hand at quarter end
X={} ##Excess Cash

##Creating the production,Quantity sold, and end of period inventory varia

bles
for i in np.arange(4):
P_a[i] = model.addVar(lb=0)
P_b[i] = model.addVar(lb=0)
Q_a[i] = model.addVar(lb=0)
Q_b[i] = model.addVar(lb=0)
I_a[i] = model.addVar(lb=0)
I_b[i] = model.addVar(lb=0)
M_a[i] = model.addVar(lb=0)
M_b[i] = model.addVar(lb=0)
B[i] = model.addVar(lb=0)
C[i] = model.addVar(lb=0)
X[i] = model.addVar(lb=0)

##Creating the constraints for production

con_production ={}
##For IBM
for i in np.arange(4):
con_production[i]=model.addConstr(P_a[i]<=prod_capacity_a[i])

##For Macintosh
for i in np.arange(4):
con_production[i+4]=model.addConstr(P_b[i]<=prod_capacity_b[i])

##Creating the constraints for end of period inventory

con_inventory={}
for i in np.arange(4):
if i==0:
con_inventory[i]=model.addConstr(I_a[i]==10000+P_a[i]-Q_a[i])##Qua
rter 1 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==5000+P_b[i]-Q_b[i])##Qu
arter 1 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
elif i<3:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 2-3 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 2-3 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 23/28

3/17/2020 Case_Study_Worksheet - final 1.0

con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
else:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 4 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 4 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=2*Q_a[i])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=2*Q_b[i])

##Creating the constraints for Quantity sold

con_sales={}
##For IBM
for i in np.arange(4):
con_sales[i]=model.addConstr(Q_a[i]==base_sales_a[i]+M_a[i]/10) ##For
IBM
con_sales[i+4]=model.addConstr(Q_b[i]==base_sales_b[i]+M_b[i]/10) ## F
or Macintosh

##Creating the constraints for the loan

con_loan={}
for i in np.arange(4):
con_loan[i]=model.addConstr(B[i]<=750000)

##Creating constraints for the cashflow

con_cashflow={}
con_cashflow={}
con_cashflow[0]=model.addConstr(C[0]==B[0]+(175*0.75*Q_a[0] + 175*0.2*8000
+ 175*0.03*7000 + 175*0.02*0.5*7000)
+(200*0.8*Q_b[0] + 200*0.11*1600 + 200*0.05*140
0 + 200*0.04*0.5*1400)
-(100*P_a[0])-(110*P_b[0])-(100*0.03*I_a[0])-(1
10*0.03*I_b[0])
-M_a[0] - M_b[0]-500000)
con_cashflow[1]=model.addConstr(C[1]==B[1]+(175*0.75*Q_a[1] + 175*0.2*Q_a[
0] + 175*0.03*8000 + 175*0.02*0.5*8000)
+(200*0.8*Q_b[1] + 200*0.11*Q_b[0] + 200*0.05*1
600 + 200*0.04*0.5*1600)
-(100*P_a[1])-(110*P_b[1])-(100*0.03*I_a[1])-(1
10*0.03*I_b[1])
-M_a[1] - M_b[1] - B[0]*(1+interest)-500000)
con_cashflow[2]=model.addConstr(C[2]==B[2]+(175*0.75*Q_a[2] + 175*0.2*Q_a[
1] + 175*0.03*Q_a[0] + 175*0.02*0.5*Q_b[0])
+(200*0.8*Q_b[2] + 200*0.11*Q_b[1] + 200*0.05*Q
_b[0] + 200*0.04*0.5*Q_b[0])
-(100*P_a[2])-(110*P_b[2])-(100*0.03*I_a[2])-(1
10*0.03*I_b[2])
-M_a[2] - M_b[2] - B[1]*(1+interest)-500000)
con_cashflow[3]=model.addConstr(C[3]==B[3]+(175*0.75*Q_a[3] + 175*0.2*Q_a[
2] + 175*0.03*Q_a[1] + 175*0.02*0.5*Q_b[1])
+(200*0.8*Q_b[3] + 200*0.11*Q_b[2] + 200*0.05*Q
_b[1] + 200*0.04*0.5*Q_b[1])
-(100*P_a[3])-(110*P_b[3])-(100*0.03*I_a[3])-(1
10*0.03*I_b[3])
-M_a[3] - M_b[3] - B[2]*(1+interest)-500000)
##Creating constraints for the excess cash
con_excess_cash={}
for i in np.arange(4):

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 24/28

3/17/2020 Case_Study_Worksheet - final 1.0

if C[i]>=0:
con_excess_cash[i]=model.addConstr(X[i]==C[i])
else: con_excess_cash[i]=model.addConstr(X[i]==0)

con_negative={}
for i in np.arange(4):
con_negative[i]=model.addConstr(P_a[i]>=0)
con_negative[i+4]=model.addConstr(P_b[i]>=0)
con_negative[i+8]=model.addConstr(M_a[i]>=0)
con_negative[i+12]=model.addConstr(M_b[i]>=0)
con_negative[i+16]=model.addConstr(B[i]>=0)

obj = C[0]+C[1]+C[2]+C[3]+(X[0]*0.045+X[1]*0.03+X[2]*0.015)-(B[3]*(1+inter
est)/1.06**(1/4))+(175*0.2*Q_a[3] + 175*0.03*Q_a[2] + 175*0.02*0.5*Q_b[2])/(1.
06**(1/4))+(175*0.03*Q_a[3] + 175*0.02*0.5*Q_b[3])/(1.06**(2/4))+ (200*0.11*Q_
b[3] + 200*0.05*Q_b[2] + 200*0.04*0.5*Q_b[2])/(1.06**(1/4))+ (200*0.05*Q_b[3]
+ 200*0.04*0.5*Q_b[3])/(1.06**(2/4))
model.setObjective(obj, GRB.MAXIMIZE)

model.optimize()

return model.objVal

In [232]: annual_interest = 0.24

quarterly_interest = annual_interest/4
sensitivity_interest(quarterly_interest)

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [1e-01, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.01s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.8976000e+32 3.000000e+30 3.897600e+02 0s
16 9.3764381e+06 0.000000e+00 0.000000e+00 0s

Solved in 16 iterations and 0.02 seconds

Optimal objective 9.376438110e+06
Out[232]: 9376438.11041825

Question 10
Minimum Expected Sales Sensitivity

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 25/28

3/17/2020 Case_Study_Worksheet - final 1.0

In [235]: def sensitivity_base_sales(error_IBM,error_MAC):

model = Model('Base Sales Sensitivity Model')

base_sales_a = [9000-error_IBM,10000-error_IBM,11000-error_IBM,12000-error
_IBM]
base_sales_b = [1800-error_MAC,2000-error_MAC,2200-error_MAC,2400-error_MA
C]

P_a={} ##Production for IBM

P_b={} ##Production for Macintosh
Q_a={} ##Quantity sold for IBM
Q_b={} ##Quantity sold for Mac
I_a={} ##End of period inventory for IBM
I_b={} ##End of period inventory for Macintosh
M_a={} ##Marketing expense for IBM
M_b={} ##Marketing expense for Macintosh
B={} ##Loan
C={} ##Cash on hand at quarter end
X={} ##Excess Cash

##Creating the production,Quantity sold, and end of period inventory varia

bles
for i in np.arange(4):
P_a[i] = model.addVar(lb=0)
P_b[i] = model.addVar(lb=0)
Q_a[i] = model.addVar(lb=0)
Q_b[i] = model.addVar(lb=0)
I_a[i] = model.addVar(lb=0)
I_b[i] = model.addVar(lb=0)
M_a[i] = model.addVar(lb=0)
M_b[i] = model.addVar(lb=0)
B[i] = model.addVar(lb=0)
C[i] = model.addVar(lb=0)
X[i] = model.addVar(lb=0)

##Creating the constraints for production

con_production ={}
##For IBM
for i in np.arange(4):
con_production[i]=model.addConstr(P_a[i]<=prod_capacity_a[i])

##For Macintosh
for i in np.arange(4):
con_production[i+4]=model.addConstr(P_b[i]<=prod_capacity_b[i])

##Creating the constraints for end of period inventory

con_inventory={}
for i in np.arange(4):
if i==0:
con_inventory[i]=model.addConstr(I_a[i]==10000+P_a[i]-Q_a[i])##Qua
rter 1 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==5000+P_b[i]-Q_b[i])##Qu
arter 1 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
elif i<3:

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 26/28

3/17/2020 Case_Study_Worksheet - final 1.0

con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 2-3 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 2-3 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=1.1*Q_a[i+1])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=1.1*Q_b[i+1])
else:
con_inventory[i]=model.addConstr(I_a[i]==I_a[i-1]+P_a[i]-Q_a[i])##
Quater 4 for IBM
con_inventory[i+4]=model.addConstr(I_b[i]==I_b[i-1]+P_b[i]-Q_b[i])
##Quarter 4 for MAC
con_inventory[i+8]=model.addConstr(I_a[i]+P_a[i]>=2*Q_a[i])
con_inventory[i+12]=model.addConstr(I_b[i]+P_b[i]>=2*Q_b[i])

##Creating the constraints for Quantity sold

con_sales={}
##For IBM
for i in np.arange(4):
con_sales[i]=model.addConstr(Q_a[i]==base_sales_a[i]+M_a[i]*0.1) ##For
IBM
con_sales[i+4]=model.addConstr(Q_b[i]==base_sales_b[i]+M_b[i]*0.1) ##
For Macintosh

##Creating the constraints for the loan

con_loan={}
for i in np.arange(4):
con_loan[i]=model.addConstr(B[i]<=750000)

##Creating constraints for the cashflow

con_cashflow={}
con_cashflow[0]=model.addConstr(C[0]==B[0]+(175*0.75*Q_a[0] + 175*0.2*8000
+ 175*0.03*7000 + 175*0.02*0.5*7000)
+(200*0.8*Q_b[0] + 200*0.11*1600 + 200*0.05
*1400 + 200*0.04*0.5*1400)
-(100*P_a[0])-(110*P_b[0])-(100*0.03*I_a[0
])-(110*0.03*I_b[0])
-M_a[0] - M_b[0]-500000)
con_cashflow[1]=model.addConstr(C[1]==B[1]+(175*0.75*Q_a[1] + 175*0.2*Q_a[
0] + 175*0.03*8000 + 175*0.02*0.5*8000)
+(200*0.8*Q_b[1] + 200*0.11*Q_b[0] + 200*0.
05*1600 + 200*0.04*0.5*1600)
-(100*P_a[1])-(110*P_b[1])-(100*0.03*I_a[1
])-(110*0.03*I_b[1])
-M_a[1] - M_b[1] - B[0]*1.04-500000)
con_cashflow[2]=model.addConstr(C[2]==B[2]+(175*0.75*Q_a[2] + 175*0.2*Q_a[
1] + 175*0.03*Q_a[0] + 175*0.02*0.5*Q_b[0])
+(200*0.8*Q_b[2] + 200*0.11*Q_b[1] + 200*0.
05*Q_b[0] + 200*0.04*0.5*Q_b[0])
-(100*P_a[2])-(110*P_b[2])-(100*0.03*I_a[2
])-(110*0.03*I_b[2])
-M_a[2] - M_b[2] - B[1]*1.04-500000)
con_cashflow[3]=model.addConstr(C[3]==B[3]+(175*0.75*Q_a[3] + 175*0.2*Q_a[
2] + 175*0.03*Q_a[1] + 175*0.02*0.5*Q_b[1])
+(200*0.8*Q_b[3] + 200*0.11*Q_b[2] + 200*0.
05*Q_b[1] + 200*0.04*0.5*Q_b[1])
-(100*P_a[3])-(110*P_b[3])-(100*0.03*I_a[3
])-(110*0.03*I_b[3])

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 27/28

3/17/2020 Case_Study_Worksheet - final 1.0

-M_a[3] - M_b[3] - B[2]*1.04-500000)

##Creating constraints for the excess cash

con_excess_cash={}
for i in np.arange(4):
if C[i]>=0:
con_excess_cash[i]=model.addConstr(X[i]==C[i])
else: con_excess_cash[i]=model.addConstr(X[i]==0)

con_negative={}
for i in np.arange(4):
con_negative[i]=model.addConstr(P_a[i]>=0)
con_negative[i+4]=model.addConstr(P_b[i]>=0)
con_negative[i+8]=model.addConstr(M_a[i]>=0)
con_negative[i+12]=model.addConstr(M_b[i]>=0)
con_negative[i+16]=model.addConstr(B[i]>=0)

obj = C[0]+C[1]+C[2]+C[3]+(X[0]*0.045+X[1]*0.03+X[2]*0.015)-(B[3]*1.04/1.0
6**(1/4))+(175*0.2*Q_a[3] + 175*0.03*Q_a[2] + 175*0.02*0.5*Q_b[2])/(1.06**(1/4
))+(175*0.03*Q_a[3] + 175*0.02*0.5*Q_b[3])/(1.06**(2/4))+ (200*0.11*Q_b[3] + 2
00*0.05*Q_b[2] + 200*0.04*0.5*Q_b[2])/(1.06**(1/4))+ (200*0.05*Q_b[3] + 200*0.
04*0.5*Q_b[3])/(1.06**(2/4))
model.setObjective(obj, GRB.MAXIMIZE)

model.optimize()

return model.objVal

In [236]: sensitivity_base_sales(1000,200)

Optimize a model with 64 rows, 44 columns and 163 nonzeros

Coefficient statistics:
Matrix range [1e-01, 2e+02]
Objective range [1e-02, 4e+01]
Bounds range [0e+00, 0e+00]
RHS range [2e+03, 8e+05]
Presolve removed 44 rows and 12 columns
Presolve time: 0.03s
Presolved: 20 rows, 32 columns, 99 nonzeros

Iteration Objective Primal Inf. Dual Inf. Time

0 3.8976000e+32 3.000000e+30 3.897600e+02 0s
16 9.3273581e+06 0.000000e+00 0.000000e+00 0s

Solved in 16 iterations and 0.05 seconds

Optimal objective 9.327358110e+06
Out[236]: 9327358.11041825

localhost:8888/nbconvert/html/Case_Study_Worksheet - final 1.0 .ipynb?download=false 28/28