Wyndor Product Mix Problem

• 2 new products: Doors and Windows.

• 3 plants
– Each Door needs: 1 hour on Plant 1 and 3 hours on Plant 3.
– Each Window needs: 2 hours on Plant 2 and 2 hours on Plant 3.

• Unit profits are: $300 and $500 for Doors and Windows, respectively.

• Hours available per week are: 4, 12, and 18 for Plants 1, 2 and 3, respectively.

Question:
What should the product mix be (i.e., how many Doors and how many Windows should be
produced per week)?
 Tabular Representation of Data

DoorsDoors Windows
Windows

Unit ProfitUnit Profit $300

$600
$500
$300

Hours
Hours Used Per Unit Produced Available
Hours Used Per Unit Produced Hours Available
Plant 1 1 0 4
Plant 1 1 0 4
Plant 2 0 2 12
Plant 2 0 2 12
Plant 3 3 2 18
Plant 3 3 2 18
Algebraic Model for Wyndor Glass Co.

Let D = number of Doors to produce

W = number of Windows to produce

Maximize P = $300D + $500W

subject to
D≤4
2W ≤ 12
3D + 2W ≤ 18
and
D ≥ 0, W ≥ 0.
 Graphing Product Mix

Production rate (units per week) for windows

A product mix of
7 D = 4 and W = 6

6 (4, 6)

4 A product mix of
D = 2 and W = 3
3 (2, 3)

1
Origin

-2 -1 0 1 2 3 4 5 6 7 8 D
Production rate (units per week) for doors
-1

-2
Graph Showing Constraints: D ≥ 0 and W ≥ 0

6
Production rate for windows

0 2 4 6 8 D
Production rate for doors
Graph Nonnegative Solutions Permitted by D ≤ 4

D=4

6
Production rate for windows

0 2 4 6 8 D
Production rate for doors
Graph Nonnegative Solutions Permitted by 2W ≤ 12

Production rate for windows

W

2 W = 12
6

0 2 4 6 8 D
Production rate for doors
Graph Nonnegative Solutions Permitted by
3D + 2W ≤ 18
Production rate for windows
W
10

3D + 2W = 18
4

0 2 4 6 8 D
Production rate for doors
 Graph Feasible Region

Production rate for windows

W
10

3 D + 2 W = 18
8

D=4

6 2 W =12

Feasible

region
2

0 2 4 6 8 D
Production rate for doors
 Graph Objective Function (P = 1,500)

W
Production rate

for windows
8

Feasible
4
region
P = 1500 = 300D + 500W

0 2 4 6 8 D
Production rate for doors
 Finding Optimal Solution
Production rate W
for windows
8

P = 3600 = 300D + 500W

Optimal solution

P = 3000 = 300D + 500W 6 (2, 6)

Feasible
4
region
P = 1500 = 300D + 500W

0 2 4 6 8 10 D
Production rate for doors

Calculate optimal solution exactly by solving for intersection

of two lines; i.e., solving 2 simultaneous equations in 2
unknowns.
 Alternative Way of Finding Optimal Solution

Clearly, optimal solution lies at one of the corners of feasible region.

Hence, to find optimal solution:

1. Calculate coordinates of each corner point (solve for intersection of two lines;
i.e., solution of 2 simultaneous linear equations in 2 unknowns)

2. For each corner point, find value of objective function (substitute coordinates
of corner point into objective function).

3. Optimal solution is corner that corresponds to optimal value of objective

function.
 Summary of Graphical Method
• Draw “constraint boundary line” for each constraint. Use origin (or any point
not on line) to determine which side of line is permitted by constraint.

• Find feasible region by determining where all constraints are satisfied

simultaneously.

• Determine slope of one objective function line. All other objective function
lines will have same slope.

• Move a straight edge with this slope through feasible region in direction of
improving values of objective function. Stop at last instant that straight edge
still passes through a point in feasible region. This line given by straight edge
is optimal objective function line. A feasible point on optimal objective
function line is an optimal solution; or, alternatively: Solve for and evaluate
corner points.
 Blue Ridge Hot Tubs
Problem
Blue Ridge would like to decide Resource required for
Resources
how many Aqua-Spa hot tubs and  
Aqua-Spa Hydro-Lux available
Hydro-Lux hot tubs to produce to
maximize the total profit. Resource Pumps 1 1 200
requirements, resource availability, Labour (hours) 9 6 1566
Tubing (feet) 12 16 2880
and unit profits are shown in the
Unit profit ($) 350 300  
table.
Model
Variables:
  = # of Aqua-Spa hot tubs to produce
= # of Hydro-Lux hot tubs to produce
Objective: MAX (total profit in $)
Constraints: S.T. (pumps)
(labor hours)
(feet of tubing)
Bounds:
 Blue Ridge Hot Tubs – graphical method
 = # of Aqua-Spa hot tubs to produce
= # of Hydro-Lux hot tubs to produce
 𝑥2 MAX (total profit in $)
S.T. (pumps)
250 (labour hours)
(feet of tubing)
200
tub
150 ing
pu
m
100 ps

 𝑥1 ≥ 0
50
lab

 𝑥2 ≥ 0
ou

0
r

0 50 100 150 200 250  𝑥1

 Blue Ridge Hot Tubs – graphical method
 = # of Aqua-Spa hot tubs to produce
= # of Hydro-Lux hot tubs to produce
 𝑥2 MAX (total profit in $)
S.T. (pumps)
250 (labour hours)
(feet of tubing)
200
tub
150 ing
pu
m
100 ps

50
lab
ou

0
r

0 50 100 150 200 250  𝑥1

 Blue Ridge Hot Tubs – graphical method – level curves
 = # of Aqua-Spa hot tubs to produce
= # of Hydro-Lux hot tubs to produce
 𝑥2 MAX (total profit in $)
S.T. (pumps)
250 (labour hours)
(feet of tubing)
200
tub Optimal
  solution = = (122,78) [found by solving
150 ing the two simultaneous equations representing the
two intersecting lines]
pu
m
100 ps
(122,78) Optimal objective function value = total profit =
= = 350(122) + 300(78) = 66100
30

50 Interpretation: optimal solution is to produce 122 Aqua-

00

Spa hot tubs and 78 Hydro-Lux hot tubs. This will give us
10

lab
0

a profit of $66,100.
00

ou

0
0

0 50 100 150 200 250  𝑥1

 Blue Ridge Hot Tubs – graphical method – corner points
 = # of Aqua-Spa hot tubs to produce
= # of Hydro-Lux hot tubs to produce
 𝑥2 MAX (total profit in $)
S.T. (pumps)
250 (labour hours)
(feet of tubing)
200(0,180)54000
The highest value is achieved by
tub
ing (80,120)64000 corner point (122,78), so (122,78)
150 is the optimal solution.
pu
m Optimal objective function value =
100 ps (122,78)66100  MAX = total profit = 66100

50 Interpretation: the optimal solution

is to produce 122 Aqua-Spa hot
lab

(0,0)0 (174,0)60900 tubs and 78 Hydro-Lux hot tubs.

ou

0 This will give us a profit of

r

0 50 100 150 200 250  𝑥1 $66,100.

Special Cases:
 = # of Aqua-Spa hot tubs to produce
= # of Hydro-Lux hot tubs to produce
MAX (total profit in $)
S.T. (pumps)
(labour hours)
(feet of tubing)
 Special Cases: Multiple optimal solutions
 = # of Aqua-Spa hot tubs to produce
= # of Hydro-Lux hot tubs to produce
 𝑥2 MAX (total profit in $)
S.T. (pumps)
250 (labour hours)
(feet of tubing)
200
tub
150 ing
pu
m
100 ps

50
lab
ou

0
r

0 50 100 150 200 250  𝑥1

 Special Cases: Multiple optimal solutions
 = # of Aqua-Spa hot tubs to produce
= # of Hydro-Lux hot tubs to produce
 𝑥2 MAX (total profit in $)
S.T. (pumps)
250 (labour hours)
(feet of tubing)
200 Optimal solutions are:
(80,120), (122,78) and all points
tub
ing(80,120)60000  MAX on segment between them.
150
pu Optimal objective function value is
m
100 ps 60000. It is achieved by each of
(122,78)60000  MAX the multiple optimal solutions.
30
50 15 000 In practice, to choose from the
00
lab

multiple optimal solutions, we

0
ou

might want to look for other

r

0 objectives or preferences.
0 50 100 150 200 250  𝑥1
 Special Cases: Infeasible Problem
Problem Model
Matt makes small and large desks. He has 12m2  = # of small desks to make
of wooden board and needs 1.5m2 for each = # of large desks to make
small desk and 3m2 for each large desk. Profit is
MAX (profit in $)
$50 per small desk and $90 per large desk. He
S.T. (m2)
already promised desks for some customers, so
needs to make at least 5 small desks and 3 large
desks. Formulate Matt’s problem.
𝑥2
  Feasible region is empty: no
6 solution satisfies all constraints.

 𝑥1 ≥5 Problem is infeasible: optimal

4 solution does not exists.
m2  𝑥2 ≥3
of b Interpretation: Matt has to get more
2 oar wooden board or he will not be able
d to make the desks that he already
promised to make.
0
0 2 4 6 8 10  𝑥1
 Special Cases:
Problem Model
A farmer would like to decide how many  = # of acres of
acres of his field to plant with watermelons watermelons
and how many with cantaloupes. The profit = # of acres of
is $25 per acre of watermelons and $40 per cantaloupes
acre of cantaloupes. The farmer can sell all MAX (profit in $)
production. However, he would like to have S.T.
at least as many acres of watermelons as
cantaloupes.
 Special Cases: Unbounded Problem
Problem Model
A farmer would like to decide how many acres of his  == ## of acres of watermelons
of acres of cantaloupes
field to plant with watermelons and how many with
cantaloupes. The profit is $25 per acre of MAX (profit in $)
S.T.
watermelons and $40 per acre of cantaloupes. The
farmer can sell all production. However, he would
like to have at least as many acres of watermelons as
cantaloupes.
𝑥2
  Problem
  is unbounded: total profit can be
15 increased without limit. Optimal solution does not
0 exists: for any feasible solution, we can always
show a better feasible solution.
10
0 Unbounded problem usually indicates that the
model is incorrect or incomplete. In this case, the
40
50 00 farmer probably forgot to say how large his field
20 is. For example, if he has 150 acres in total, we
00 would add a constraint and we would find an
0 𝑥 optimal solution.
0 50 100 150   1
 Develop a Spreadsheet Model
• Step #1: Data Cells
– Enter all of data for problem on spreadsheet.
– Make consistent use of rows and columns.
– Good idea: colour code “data cells” (e.g., light blue).

• Step #2: Changing Cells

– Add a cell in spreadsheet for every decision that needs to be made.
– Good idea: colour code “changing cells” (e.g., yellow).

• Step #3: Objective Cell

– Develop an equation that defines objective of model.
– Typically, this equation involves data cells and changing cells to determine a quantity of interest (e.g., total profit
or total cost).
– Good idea: colour code this cell (e.g., orange).

• Step #4: Constraints

– For any restricted resource, calculate amount used in a cell on spreadsheet (an output cell).
– Define constraint in three consecutive cells. For example, if Quantity A <= Quantity B, put these 3 items
(Quantity A, <=, Quantity B) in consecutive cells.

• Step #5: Trial Solution

– Enter trial solution to check model
 Develop a Spreadsheet Model

Please refer to spreadsheet provided (Wyndor.xlsx)

 Using Solver
Identifying Objective Cell and Changing Cells
• Choose “Solver” from Data tab.
• Select cell you wish to optimize in “Set Objective” window.
• Choose “Max” or “Min” depending on whether you want to maximize or minimize
objective cell.
• Enter all changing cells in “By Changing Variable Cells” window.

Doors Windows
Unit Profit $300 $500
Hours Hours
Hours Used Per Unit Produced Used Available
Plant 1 1 0 1 <= 4
Plant 2 0 2 2 <= 12
Plant 3 3 2 5 <= 18

Doors Windows Total Profit

Units Produced 1 1 $800
 Adding Constraints

• To begin entering constraints, click “Add” button to right of constraints

window.
• Fill in entries in resulting Add Constraint dialogue box.
Doors Windows
Unit Profit $300 $500
Hours Hours
Hours Used Per Unit Produced Used Available
Plant 1 1 0 1 <= 4
Plant 2 0 2 2 <= 12
Plant 3 3 2 5 <= 18

Doors Windows Total Profit

Units Produced 1 1 $800
Complete Solver Dialogue Box
 Optimal Solution

Doors Windows
Unit Profit $300 $500
Hours Hours
Hours Used Per Unit Produced Used Available
Plant 1 1 0 2 <= 4
Plant 2 0 2 12 <= 12
Plant 3 3 2 18 <= 18

Doors Windows Total Profit

Units Produced 2 6 $3,600
 Profit & Gambit Co.
• Management has decided to undertake a major advertising campaign, using both
television and print media, to increase sales of three key products:
– A spray prewash stain remover (S).
– A liquid laundry detergent (L).
– A powder laundry detergent (P).

• Management has set following goals for campaign:

– Sales of S, L and P should increase by at least 3%, 18%, and 4%, respectively.

  TV Print      
Unit Cost ($millions) 1 2      
           
  Increase in Sales per Unit of Advertising     Minimum
Increase
S 0% 1%     3%
L 3% 2%     18%
P -1% 4%     4%
Question: how much should they advertise in each medium to meet sales goals at a
minimum total cost?
 Algebraic Model for Profit & Gambit

Let TV = number of units of advertising on television

PM = number of units of advertising in print media

Minimize Cost = TV + 2PM (in millions of dollars)

subject to
Stain remover increased sales: PM ≥ 3
Liquid detergent increased sales: 3TV + 2PM ≥ 18
Powder detergent increased sales: –TV + 4PM ≥ 4
and
TV ≥ 0, PM ≥ 0.
 Profit & Gambit Co. Spreadsheet Model

Television Print Media

Unit Cost ($millions) 1 2

Increased Minimum
Increase in Sales per Unit of Advertising Sales Increase
Stain Remover 0% 1% 3% >= 3%
Liquid Detergent 3% 2% 18% >= 18%
Powder Detergent -1% 4% 8% >= 4%

Total Cost
Television Print Media ($millions)
Advertising Units 4 3 10
 Applying Graphical Method
Amount of print media advertising
PM

Feasible
10
region

PM = 3
2
-TV + 4 PM = 4
3 TV + 2 PM = 18

-4 -2 0 2 4 6 8 10 TV
Amount of TV advertising
 Optimal Solution
PM

10
Feasible
Cost = 15 = TV + 2 PM region

Cost = 10 = TV + 2 PM

4
(4,3)
optimal
solution
0 5 10 15 TV
Amount of TV adverti sing

Feasible region is unbounded. Yet, problem itself is bounded and

has an optimal solution; i.e., whether the problem is unbounded or
not depends on both the feasible region and the objective function.
 Summary

• In
  LP problems, there are normally many solutions to choose from (feasible region).
• Number of optimal solutions of an LP problem may be:
0 Zero (when the problem is infeasible or unbounded)
1 One (always in one of the corner points)
Infinite (when multiple solutions are optimal)
No other possibility exists. For example, it is not possible to have exactly two
optimal solutions.
• If an LP problem has optimal solutions (is neither infeasible nor unbounded), then:
– All optimal solutions are on the boundary of the feasible region.
– At least one corner point is an optimal solution.

• An LP problem may be unbounded only if the feasible region is unbounded. However,

not every problem with unbounded feasible region is unbounded (some have optimal
solutions; depends on the objective function).

• All above statements hold for LP optimization problems, but not necessarily for other
optimization problems such as NLP, ILP, MOLP.
 Additional remarks
•  Signs and are not permitted in constraints. Only , , and are permitted.
– Example: What would be the optimal solution of: MAX S.T. ?

• MAX is equivalent to MIN

– Example: MAX Profit is equivalent to MIN Loss (Loss = – Profit)

• MAX positive constant is equivalent to MAX (similarly for MIN)

• MAX is equivalent to MAX (similarly for MIN)

• Graphical method is not practical

(only for problems with at most two variables),
but it is useful for learning & gaining insights.

• Example of a three-variable feasible region 

 Recap
• LP Model: The mathematical model that represents a LP problem.
• Parameter: The parameters of a LP model are the constants (coefficients or right-hand sides) in the
functional constraints and the objective function. Each parameter represents a quantity (e.g., the amount
available of a resource) that is of importance for the analysis of the problem.
• Non-negativity Constraint: A constraint that expresses the restriction that a particular decision variable
must be nonnegative (greater than or equal to zero).
• Functional Constraint: A constraint with a function of the decision variables on the left-hand side. All
constraints in a LP model that are not non-negativity constraints are called functional constraints
• Feasible Solution: A solution that simultaneously satisfies all the constraints in a mathematical
programming model.
• Feasible Region: The geometric region that consists of all the feasible solutions.
• Infeasible Solution: A solution that violates at least one of the constraints in the LP model.
• Objective Function: The part of a LP model that expresses what needs to be either maximized or
minimized, depending on the objective for the problem.
• Optimal Solution: The best feasible solution according to the objective function.
• Unbounded Problem: A problem where constraints and objective function are such that the objective
function value could be increased (for maximization) or decreased (for minimization) ad-infinitum.