Problem Set C (Sensitivity Analysis)
Note: Dual price and shadow prices are same.
Problem 1
Consider the following linear program:
Max 3A + 2B
s.t.
A + B <= 10
3A + B <= 24
A + 2B <= 16
A, B >= 0
a. Use the graphical solution procedure to find the optimal solution. [(7, 3), 27]
b. Assume that the objective function coefficient for A changes from 3 to 5. Does the
optimal solution change? Use the graphical solution procedure to find the new optimal
solution. [No, 41]
c. Assume that the objective function coefficient for A remains 3, but the objective
function coefficient for B changes from 2 to 4. Does the optimal solution change? Use
the graphical solution procedure to find the new optimal solution. [Yes, (4,6), 36]
d. The computer solution for the linear program in part (a) provides the following
objective coefficient range information:
Variable Objective Allowable Allowable
coefficients increase decrease
A 3.00000 3.00000 1.00000
B 2.00000 1.00000 1.00000
Use this objective coefficient range information to answer parts (b) and(c).
Problem 2
Consider the linear program in Problem 1. The value of the optimal solution is 27. Suppose
that the right-hand side for constraint 1 is increased from 10 to 11.
a. Use the graphical solution procedure to find the new optimal solution. [(6.5, 4.5), 28.5]
b. Use the solution to part (a) to determine the shadow price (dual value) for constraint 1.
[1.5]
c. The sensitivity report for the linear program in Problem 1 provides the following right-
hand-side range information:
Constraint RHS Value Allowable Increase Allowable Decrease
1 10.000 1.20000 2.00000
2 24.000 6.00000 6.00000
3 16.000 Infinite 3.00000
What does the right-hand-side range information for constraint 1 tell you about the
shadow price for constraint 1?
d. The shadow price for constraint 2 is 0.5. Using this shadow price and the right-hand-
side range information in part (c), what conclusion can you draw about the effect of
changes to the right-hand side of constraint 2?
Problem 3
Investment Advisors, Inc., is a brokerage firm that manages stock portfolios for a number
of clients. A particular portfolio consists of U shares of U.S. Oil and H shares of Huber
Steel. The annual return for U.S. Oil is $3 per share and the annual return for Huber Steel is
$5 per share. U.S. Oil sells for $25 per share and Huber Steel sells for $50 per share. The
portfolio has $80,000 to be invested. The portfolio risk index (0.50 per share of U.S. Oil and
0.25 per share for Huber Steel) has a maximum of 700. In addition, the portfolio is limited
to a maximum of 1000 shares of U.S. Oil. The linear programming formulation that will
maximize the total annual return of the portfolio is as follows:
Max 3U + 5H
s.t.
25U + 50H <= 80,000 Funds Available
0.50U + 0.25H <= 700 Risk Maximum
U <= 1000 U.S. Oil maximum
U, H >= 0
The computer solution of this problem is shown below:
Final Reduce Objective Allowabl Allowabl
Cel Name Value Cost Coefficient Increase Decrease
$A$ U S Oil 800 0 3 7 0.5
$B$ HUBER STL 1200 0 5 1 3.5
Final Shadow Constraint Allowabl Allowabl
Cel Name Value Price R.H. Side Increase Decrease
$D$ Funds Available 80000 0.0933 80000 60000 15000
$D$ Risk Maximum 700 1.3333 700 75 300
$D$ U.S. Oil maximum 800 0 1000 1E+30 200
a. What is the optimal solution, and what is the value of the total annual return? [(800,
1200), $8400]
b. Which constraints are binding? What is your interpretation of these constraints in
terms of the problem? [constraints 1 & 2]
c. What are the dual values for the constraints? Interpret each. [0.09333, 1.3333, 0]
d. Would it be beneficial to increase the maximum amount invested in U.S. Oil? Why or
why not? [no, already unutilized]
e. How much would the return for U.S. Oil have to increase before it would be
beneficial to increase the investment in this stock? [more than $7 per share]
f. How much would the return for Huber Steel have to decrease before it would be
beneficial to reduce the investment in this stock? [more than $3.5 per share]
g. How much would the total annual return be reduced if the U.S. Oil maximum were
reduced to 900 shares? [no reduction]
Problem 4
Quality Air Conditioning manufactures three home air conditioners: an economy model, a
standard model, and a deluxe model. The profits per unit are $63, $95, and $135,
respectively. The production requirements per unit are as follows:
Number of Number of Manufacturing
Fan motors Cooling Coils time (Hours)
Economy o
1 1 8
Standard 1 2 12
Deluxe 1 4 14
For the coming production period, the company has 200 fan motors, 320 cooling coils, and
2400 hours of manufacturing time available. How many economy models (E), standard
models (S), and deluxe models (D) should the company produce in order to maximize profit?
a. What is the optimal solution, and what is the value of the objective
function? [(80,120,0), 16440]
b. Which constraints are binding? [constraints 1 & 2]
c. Which constraint shows extra capacity? How much? [constraint 3, 320]
d. If the profit for the deluxe model were increased to $150 per unit, would the optimal
solution change? [no]
e. Identify the range of optimality for each objective function coefficient. [47.5-75, 87-
126, -1E+30 - 159],
f. Suppose the profit of the economy model is increased by $6 per unit, and the profit
for the standard model is decreased by $2 per unit, and the profit for the deluxe model
is increased by $4 per unit. What will the new optimal solution be? [(80,120,0),
16680]
g. Identify the range of feasibility for the right-hand-side values. [160-280, 200-400,
2080- 1E+30]
h. If the number of fan motors available for production is increased by 100, will the dual
value for that constraint change? Explain. [yes]
Problem 5
Digital Controls, Inc. (DCI) manufactures two models of a radar gun used by police to
monitor the speed of automobiles. For the next week, the company has orders for 100 units of
model A and 150 units of model B. DCI wants to just meet the demand. Although DCI
purchases all the electronic components used in both models, the plastic cases for both
models are manufactured at a DCI plant. Each model A case requires 4 minutes of injection-
molding time and 6 minutes of assembly time. Each model B case requires 3 minutes of
injection-molding time and 8 minutes of assembly time. For next week the Newark plant has
600 minutes of injection-molding time available and 1080 minutes of assembly time
available. The manufacturing cost is $10 per case for model A and $6 per case for model B.
Depending upon demand and the time available at the Newark plant, DCI occasionally
purchases cases for one or both models from an outside supplier in order to fill customer
orders that could not be filled otherwise. The purchase cost is $14 for each model A case and
$9 for each model B case. Management wants to develop a minimum cost plan that will
determine how many cases of each model should be produced at the Newark plant and how
many cases of each model should be purchased.
a. What is the optimal solution and what is the optimal value of the objective function?
[Manufacture A:100, B: 60, Purchase A: 0, B: 90, $2170]
b. Which constraints are binding? [assembly time, demand of A and demand off B]
c. What are the dual prices? Interpret each. [injection molding time: 0, assembly time: -
0.375, demand A: 12.25, demand B: 9]
d. If you could change the right-hand side of one constraint by one unit, which one
would you choose? Why? [assembly time]
e. Interpret the range of optimality for the objective function coefficients.
[manufacturing cost A: - 1E+30 – 11.75, manufacturing cost B: 3.6666-9, Purchasing
cost A: 12.25-1E+30, purchasing cost B: 6-11.3333]
f. Suppose that the manufacturing cost increases to $11.20 per case for model A. What
is the new optimal solution? [remains same, $2290]
g. Suppose that the manufacturing cost increases to $11.20 per case for model A and the
manufacturing cost for model B decreases to $5 per unit. Would the optimal solution
change? [Re-solve the model, Manufacture A: 0, Manufacture B: 135, Purchase A:
100 Purchase B: 15, $2210]
Problem 6
Benson Electronics manufactures three components used to produce cell telephones and other
communication devices. In a given production period, demand for the three components may
exceed Benson’s manufacturing capacity. In this case, the company meets demand by
purchasing the components from another manufacturer at an increased cost per unit. Benson’s
manufacturing cost per unit and purchasing cost per unit for the three components are as
follows:
Source Component 1 Component 2 Component 3
Manufacture $4.50 $5.00 $2.75
Purchase $6.50 $8.80 $7.00
Manufacturing times in minutes per unit for Benson’s three departments are as follows:
Department Component 1 Component 2 Component 3
Production 2 3 4
Assembly 1 1.5 3
Testing & Packaging 1.5 2 5
For the next production period, Benson has capacities of 360 hours in the production
department, 250 hours in the assembly department, and 300 hours in the testing and
packaging department.
a. Formulate a linear programming model that can be used to determine how many
units of each component to manufacture and how many units of each component to
purchase. Assume that component demands that must be satisfied are 6000 units for
component 1, 4000 units for component 2, and 3500 units for component 3.
b. What is the optimal solution? How many units of each component should be
manufactured and how many units of each component should be purchased?
[manufacture component 1: 2000, component 2: 4000, component 3: 1400, purchase
component 1: 4000, component 2: 0, component 3: 2100, $73550
c. Which departments are limiting Benson’s manufacturing quantities? Use the dual
price to determine the value of an extra hour in each of these departments.
[production and the testing & packaging departments, $ -54.3750, $0, $-7.5]
d. Suppose that Benson had to obtain one additional unit of component 2. Discuss what
the dual price for the component 2 constraints tells us about the cost to obtain the
additional unit. [The dual value is $7.969. This tells us that the value of the optimal
solution will worsen (the cost will increase) by $7.969 for an additional unit of
component 2. Note that although component 2 has a purchase cost per unit of $8.80,
it would only cost Benson $7.969 to obtain an additional unit of component 2.
Problem 7
Round Tree Manor is a hotel that provides two types of rooms with three rental classes: Super
Saver, Deluxe, and Business. The profit per night for each type of room and rental class is as
follows:
Rental Class
Super Deluxe Business
saver
Type I $30 $35 -
Room
Type II $20 $30 $40
Ty
pe I rooms do not have Internet access and are not available for the Business rental class.
Round Tree’s management makes a forecast of the demand by rental class for each night in
the future. A linear programming model developed to maximize profit is used to determine
how many reservations to accept for each rental class. The demand forecast for a particular
night is 130 rentals in the Super Saver class, 60 rentals in the Deluxe class, and 50 rentals in
the Business class. Round Tree has 100 Type I rooms and 120 Type II rooms.
a. Use linear programming to determine how many reservations to accept in each rental
class and how the reservations should be allocated to room types. Is the demand by
any rental class not satisfied? Explain.
b. How many reservations can be accommodated in each rental class? [Super Saver
110; Deluxe: 60; Business:50]
c. Management is considering offering a free breakfast to anyone upgrading from a
Super Saver reservation to deluxe class. If the cost of the breakfast to Round Tree is
$5, should this incentive be offered? [Yes, the effect of a person upgrading is an
increase in demand for Deluxe accommodations from 60 to 61. From constraint 2, we
see that such an increase in demand will increase profit by $10. The added cost of the
breakfast is only $5.]
d. With a little work, an unused office area could be converted to a rental room. If the
conversion cost is the same for both types of rooms, would you recommend
converting the office to a Type I or a Type II room? Why? [Convert to a Type I room.
From the dual value to constraint 4 we see that this will increase profit by $30.]
e. Could the linear programming model be modified to plan for the allocation of rental
demand for the next night? What information would be needed and how would the
model change? [Yes. We would need the forecast of demand for each rental class on
the next night. Using the demand forecasts, we would modify the right-hand sides of
the first three constraints and re-solve.]
