UNIT III – OPTIMIZATION MODELS

An optimization model is a translation of the key characteristics of the business problem you are trying to solve.
The model consists of three elements: the objective function, decision variables and business constraints.

SPREADSHEETS

A good model has four key attributes:

1.      Consistency,

2.      Robustness,

3.      Flexibility

4.      Transparency.
UNIT III – OPTIMIZATION MODELS

An optimization model is a translation of the key characteristics of the business problem you are trying to solve.
The model consists of three elements: the objective function, decision variables and business constraints.

SPREADSHEETS
When doing economic analysis in business decision making, certain operations need to be repeated with different
data at different point of time. Hence it becomes difficult if done manually. Since the use of electronic spread sheets
does this job efficiently, it is necessary for one studying business administration to learn this.

A good model has four key attributes:

1.      Consistency,
Models constructed consistently are easier to understand as users become familiar with both their purpose and
content. This will in turn give users more comfort about model integrity and make it easier to add or remove
business units, categories, numbers of periods, scenarios, etc.
2.      Robustness,
Models should be materially free from error, mathematically accurate, and readily auditable. Key output sheets
should ensure that error messages cannot occur.
When building, it is often worth keeping in mind hidden assumptions in formulae. For example, a simple gross
margin calculation may calculate profit divided by sales. However, if sales are non-existent or missing, this
calculation would give #DIV/0!. The user therefore has two options:
·         Use an IF statement to check that sales are not zero (proactive test); or
·         Construct an error check to flag if sales are zero
A good model must have error checks, sensitivity checks and Alert checks.
3.      Flexibility
One benefit of modelling in a spread sheet package such as Excel is to be able to change various assumptions and
see how these adjustments affect various outputs.
Therefore, when building a model, the user should consider what inputs should be variable and how they should be
able to vary. This may force the model builder to consider how assumptions should be entered.
The most common method of data entry in practice is simply typing data into worksheet cells, but this may allow a
model’s inputs to vary outside of scoped parameters. For example, if I have a cell seeking “Volumes”, without
using data validation I could enter “3”, “-22.8”, or “dog” in that cell. Negative volumes are nonsensical, and being
able to enter text may cause formula errors throughout the model. Therefore, the user may wish to consider other
methods of entry including using drop-down boxes, option buttons, check boxes, and so on.
4.      Transparency.
Models must be clear, concise, and fit for the purpose intended. If you can follow it on a piece of paper
(ie, no Formula bar), it’s transparent.
Most Excel users are familiar with keeping inputs and assumptions away from calculations — and keeping inputs
and assumptions and calculations away from outputs. However, this concept can be extended: It can make sense to
keep different areas of a model separate, eg, revenue assumptions on a different worksheet from cost(s) of goods
sold assumptions, and capital expenditure assumptions on a third sheet, and so on. This makes it easier to reuse
worksheets and ring-fence data. Keeping base case data away from sensitive data is also important, as many
modelling mistakes have been made from users changing the wrong, yet similar, inputs.
EXERCISE 1: BUILDING A GOOD SPREADSHEET MODEL

PROBLEM Nowlin Plastics produces a line of cell phone covers. Nowlin's best selling cover is its viper mod
The annual fixed cost for the viper cover is $234, 000. This fixed cost includes management tim
of the number of units eventually produced. In addition, the total variable cost, including labour
is considering outsourcing the production of some products for next year, including the viper. N
expensive per unit to outsource, the fixed cost can be avoided if Nowlin purchases rather than m
AIM:
To build a model to solve the make or buy problem.

PROCEDURE:

STEP 1
Draw influence diagrams. An influence diagram is a visual representation that shows which entities influence
Parts of the model are represented by circular or oval symbols called nodes, and arrows connecting the nod

Difference

Total Total
Manufacturing outsource cost
cost

Fixed cost Variable cost Quantity

Required

Variable cost
per unit

Step 2 Use influence diagram to build a mathematical model

This cost is a function of the fixed cost, the variable cost per unit and the quantity required.
Let q= quantity required
FC= the fixed cost of manufacturing
VC= variable cost of manufacturing
TMC(q)= total cost to manufacture q units
P= the per unit purchase cost
TPC(q)= the total cost to purchase q units
 S(q)= the savings from outsourcing q units

TMC (q)= FC+ (VC x q)

TPC(q)= $3.5q

S(q)= TMC(q) - TPC(q)

CALCULATIONS

Nowlin Plastics

Parameters
Manufacturing fixed cost 234000
Manufacturing variable cost per unit 2

Outcourcing cost per unit 3.5

Model
Quantity 10000 assumption

Total cost to produce 254000 (=G60 + G66*G61)

Total cost to outsource 35000 (=G63*G66)

Savings due to outsourcing 219000 (=G68-G70)

RESULT:
We had set the quantity as 10000, this can be changed to see the difference in the savings due to
est selling cover is its viper model, a slim but very durable black and gray plastic cover.
d cost includes management time, advertising and other costs that are incurred regardless
variable cost, including labour and material costs, is $2 for each unit produced. Nowlin
ext year, including the viper. Nowlin has a bid from an outside firm for $3.50 per unit. Although it is more
Nowlin purchases rather than manufactures the product. The exact demand for viper for next year is not yet known.

shows which entities influence others in a model.

, and arrows connecting the nodes show influence.

otal
utsource cost

Purchase
cost per unit

uantity required.
difference in the savings due to outsourcing.
EXERCISE 2: WHAT-IF ANALYSIS

PROBLEM Nowlin plastics Make or Buy Problem

AIM To conduct What-If analysis

PROCEDURE

STEP 1 Select cells H23:I36

STEP 2 Click the DATA tab in the Ribbon

STEP 3 Click What-If Analysis in the data tools group, and select data table

STEP 4 When the data table dialog box appears, enter G31 in the column input cell : box
Click OK

CALCULATIONS
Nowlin Plastics

Parameters Quantity 219000

Manufacturing fixed cost 234000 0
Manufacturing variable cost per unit 2 25000
50000
Outcourcing cost per unit 3.5 75000
100000
Model 125000
Quantity 10000 150000
175000
Total cost to produce 254000 200000
225000
Total cost to outsource 35000 250000
275000
Savings due to outsourcing 219000 300000

RESULT
RESULTS OF ONE-WAY DATA TABLE FOR NOWLIN PLASTICS

Nowlin Plastics

Parameters Quantity 219000

Manufacturing fixed cost 234000 0
Manufacturing variable cost per unit 2 25000
50000
Outcourcing cost per unit 3.5 75000
100000
Model 125000
Quantity 10000 150000
175000
Total cost to produce 254000 200000
225000
Total cost to outsource 35000 250000
275000
Savings due to outsourcing 219000 300000

We have quantified the savings due to outsourcing for a number of quantities.

For quantities of 150000 units or less, outsourcing is cheaper than Manufacturing and
For quantities of 175000 units or more, manufacturing is cheaper than outsourcing
acturing and
EXERCISE 3: LEARNING SOME USEFUL EXCEL FUNCTIONS FOR MODELING

AIM Learning some useful excel functions for Modeling such as SUM, SUMPRODUCT, IF, COUNTIF and VL

PROCEDURE
SUM Adds up all of the numbers in a range of cells
SUM(range)
SUMPRODUCT Returns the sum of the products of elements in a set of arrays
SUMPRODUCT(array1, array2)
IF and COUNTIF IF(condition, result if condition is true, result if condition is false)
COUNTIF(range, condition)
VLOOKUP Allows a user to pull a subset of data from a larger table of data based on some crite
VLOOKUP(value, table, index, range)

CALCULATIONS
Assign SUMPRODUCT(D21:F23,D28:F30) to calculate the transportation cost

X Y Z SUPPLY
A 8 7 3 60
Cost table B 3 8 9 70
C 11 3 5 80

DEMAND 50 80 80

0 0 60
Allocation 50 0 20
0 80 0

RESULT

Transportation cost 750

UCT, IF, COUNTIF and VLOOKUP

of data based on some criterion

e the transportation cost

Microsoft Excel 15.0 Sensitivity Report
Worksheet: [unit3-OPTIMIZATION MODELS.xlsx]Maximization- Ex4
Report Created: 29-04-2021 14:13:35

Variable Cells
Final Reduced Objective Allowable Allowable
Cell Name Value Cost Coefficient Increase Decrease
$L$29 mix 0 0 0 0 1E+030
$M$29 mix 0 0 0 0 1E+030
$N$29 mix 0 0 0 0 1E+030

Constraints
Final Shadow Constraint Allowable Allowable
Cell Name Value Price R.H. Side Increase Decrease
$O$33 constraint 1 0 0 0 1E+030 0
$O$34 constraint2 0 0 0 1E+030 0
$O$35 0 0 0 1E+030 0
$P$35 0 0 0 1E+030 0
$Q$35 0 0 0 1E+030 0
EXCERSICE 4 : Simple Maximization problem in LPP

PROBLEM: Maximize Z= 5x1 + 12x2 + 4x3

Subject to constraints
x1 + 2x2 <= 5
5x1 - x2 - 2x3 <=2
x1, x2, x3 >=0

AIM: Solving the decision variables to find the maximized profit.

PROCEDURE

STEP 1 Click the DATA tab in the Ribbon and click Solver
STEP 2 Assign Sum product to the following cells : SUMPRODUCT(C29:E29, C30:E30) and
SUMPRODUCT(C29:E29, C33:E33) SUMPRODUCT(C29:E29, C34:E34)
STEP 3 Enter C36 in the set objective box, select Max
STEP 4 Enter C28:E28 in the By changing variable cells: box
STEP 5 Click the add button to add constraints
Enter F34:F35 in the cell reference
Select <= from the drop down button
Enter G34:35 in the constraint, then click OK
STEP 6 Select Simplex LP method and click Solve
STEP 7 When the solver results dialog box appears,
Select keep solver solution, and click OK

CALCULATIONS
objective x1 x2 x3
mix 0 2.5 2.25
unit profit 5 12 4

used
constraint 1 1 2 0 5 5
constraint2 5 -1 2 2 2

RESULT
profit 39
E29, C30:E30) and
EXCERSICE 5 : Simple Minimization problem in LPP

PROBLEM: Minimize Z= 12x1 + 20x2

Subject to constraints
6x1 + 8x2 >= 100
7x1 +12x2 >=120
x1, x2 >=0

AIM: Solving the decision variables to find the minimized cost.

PROCEDURE

STEP 1 Click the DATA tab in the Ribbon and click Solver
STEP 2 Assign Suum product to the following cells : SUMPRODUCT(C30:D30, C31:D31) and
SUMPRODUCT(C30:D30, C34:D34) SUMPRODUCT(C30:D30, C35:D35)
STEP 3 Enter C38 in the set objective box, select Min
STEP 4 Enter C30:D30 in the By changing variable cells: box
STEP 5 Click the add button to add constraints
Enter E34:E35 in the cell reference
Select >= from the drop down button
Enter F34:F35 in the constraint, then click OK
STEP 6 Select Simplex LP method and click Solve
STEP 7 When the solver results dialog box appears,
Select keep solver solution, and click OK

CALCULATION

objective x1 x2
mix 15 1.25
cost 12 20

used
constraint1 6 8 100 100
constraint2 7 12 120 120

RESULT
cost 205
C31:D31) and
After solver solution box, choose sensitivity analysis option.