Lesson Plan on Basic Quantitative Techniques for Decision Making

Management Science – discipline devoted to studying and developing procedures to help the process of making
decisions and uses the scientific method for decision making.

1. Linear Programming

It is mathematical technique that permits the determination of the best or optimum use of the available resources.

This technique was first developed and applied in 1947 for the U.S. Army by George B. Dantzig, Marshall
Wood and their associates.

Linearity means that the equation or inequality must not contain quadratic components; it is tacitly assumed that
there is a single goal that can be represented by a linear objective function and that all restrictions are linear in
nature.

Programming means developing an algorithm (system of procedures) in solving the issue.

Certainty is assumed thus all parameters of the model should be known.

Applications:

1
a. Determination of the product mix to maximize contribution margin.
b. Selection of an investment mix
c. Determination of materials mix to minimize costs.
d. Assignment of jobs to machines
e. Determination of transportation routes.

Steps:

a. Formulate the objective function - maximize profit or minimize cost (peso values)

b. Formulate the constraints function - limitations of decision variables

1. Explicit constraints – apparent limitations as expressed in the problem

2. Implicit constraints – implied limitations or more commonly known as the non-negativity constraints
(the variables are not supposed to be a negative value because we cannot produce a negative number of units)

Computational Methods

a. Graphic method - involves only two variables; optimal solution is represented in corner points of the graphic
solution; solutions are generally confined to the positive quadrant.

Extreme Point Theorem states that If an optimal solution to a linear programming problem exists, then at
least one such optimal solution must be an extreme point solution.

*In a minimization problem  optimal point is the point nearest 0 taking into consideration all points.
(least cost as it approaches to 0)

*In a maximization problem  optimal point is the point farthest from 0 taking into consideration all
points. (highest revenue is reached farthest from 0)

b. Simplex method - iteration method used to reach the optimal solution; used regardless of the number of variables
involved.

*In a minimization problem  optimal solution is reached if the index row does not indicate any negative
value (must be either 0 or positve) meaning the values (cost) can no longer be decreased further.

*In a maximization problem  optimal solution is reached if the index row does not indicate any positive
value (must be either 0 or negative) meaning the values (profits) can no longer be increased further.

*Slack variable (artificial variable or dummy variable ) is introduced to equalize an inequality.

Special Types of Linear Programming:

1. Transportation Problem
2. Assignment Problem

Techniques in set up of Initial Tableau:

1. Northwest method – assign resources first to the northwest cell.

2. Row Minimum Method – assign resources first to the cell with the minimum amount (if minimization
problem) or maximum amount (maximization problem) in each row.

3. Vogel’s Approximation Method – assign resources first to the cell with the highest penalty (difference
between the least cost and next least cost or highest revenue to second highest revenue)

4. Greedy method – choose the least cost or highest revenue cell first in the whole table.

Sensitivity Analysis – involves the exploration of changes in model output in response to changes in input
parameters; it can aid in generating alternatives and in planning for responses to unpredictable changes.

1. Changes in objective function coefficients

2
 2. Changes in right hand side values
3. Adding a new variable
4. Adding a new constraint
5. Changes in the constraint coefficients

6. Shadow Pricing - associated with a constraint showing the amount of improvement in the optimal
objective function values as the value of the right hand side of that constraint is increased by one unit with
all other model parameters unchanged.

Exercise on Linear Programming

***RX produces two products, Chico and Delamar. Product Chico requires two hours of grinding and four hours of
polishing. Product Delamar requires five hours of grinding and two hours of polishing. The manufacturer has 3
grinders and 2 polishers; therefore in a 40-hour week, there are 120 hours of grinding capacity and 80 hours of
polishing capacity. There is a ready market for both products and the contribution margin per unit of Product Chico
and Product Delamar are P30 and P40 respectively.

Required: 1. Objective function

2. Constraint Functions (explicit and implicit)

*** The K-lite Cafe produces a chemical solution used for cleaning carpets. This chemical is made from a mixture
of two other chemicals which contain cleaning agent Joshua and cleaning agent Jinji. Their product must contain
175 units of agent Joshua and 150 units of Jinji and weigh at least 100 pounds. Chemical A costs P8 per pound,
while chemical B costs P6 per pound. Chemical A contains one unit of agent Joshua and three units of agent Jinji.
Chemical B contains seven units of agent Joshua and one unit of agent Jinji.

Required: Set up the linear programming model

3
 2. Learning Curve

This theory stipulates that every time the cumulative quantity of units produced is doubled, the cumulative average
time per unit is reduced by a given percentage.

Assumptions:

1. The time required to perform a given task becomes progressively shorter.

2. This technique is only applicable to the early stages of production or of any new task.
3. Ordinarily, the curve is expressed in a percentage of reduced time to complete a task for each doubling of
cumulative production.

If it is assumed that the reduction is 20%, it means that the second unit requires 80% of the cumulative average time
per unit required for the first unit; the fourth unit 80% of the second; the eighth unit 80% of the fourth and so on.
This is referred to as an 80% learning curve.

Cumulative Cumulative Cumulative Individual Unit

Number of Units Average Time Total Time Time for Nth
per unit (hours) (hours) Unit (hours)

1 100.00 100.00 100.00

2 80.00 (100 x 80%) 160.00 60.00
3 70.21 * 210.63 50.63
4 64.00 (80 x 80%) 256.00 45.37
5 59.57 * 297.85 41.85
6 56.17 * 337.02 39.17
7 53.45 * 374.15 37.13
8 51.20 (64 x 80%) 409.60 35.45

16 40.96 (51.2 x 80%) 655.36 28.06

*The 3rd, 5th, 6th etc unit are computed based on the following formula:

b
Y=aX

where Y = Cumulative average time (labor hours) per unit

X = Cumulative number of units produced
a = Time (labor hours) required to produce the first unit
b = Factor used to calculate cumulative average time to produce units

The value b is calculated as:

b= ln (learning curve % in decimal form)

ln 2

For an 80% learning curve

b= ln .80
ln 2
b= -0.3219

As an illustration, when X = 3, a = 100 and b = -0.3219

-0.3219
y = 100 x 3

y = 70.21 labor hours

Applications:
a. Setting of prices based on estimates of expected costs.
b. Scheduling labor requirements.
c. Capital budgeting decisions
d. Setting incentive wage rates.

Exercises on Learning Curve

*** WW Company plans to begin production of a new product on July 1, 200A. An 80% learning curve is
applicable to WW's manufacturing operations. If it is expected to take 1,000 direct labor hours to produce the first
unit, how many direct labor hours should it take to produce a total of eight units?
4
*** The XX Company's new process will be carried out in one department. The production process has an expected
learning curve of 80% for each doubling of output. The costs subject to the learning effect for the first batch
produced by the process were P20,000.

Required: Using the simplest learning function, the cumulative average cost per batch subject to learning effect
after the 16th batch has been produced may be estimated as ______________.

3. Expected Value

Probability – defined as the measure of how an event is most likely to happen

The probability of event A is the number of ways event A can occur divided by the total number of possible
outcomes.

a. Equal Likely to Happen

Illustration: Rolling a dice

Probability of getting a 1, 2, 3, 4, 5 or 6 in the first roll of the dice is 1/6 since all #s are given equal chances of
coming up.

Similarly, the probability of getting an odd or even number will be 50% (2, 4, 6 with 1/6 probability each and 1, 3, 5
with 1/6 probability each).

b. Outcomes are not Equally Likely to Happen

Illustration: A glass jar contains 5 red, 3 green and 2 yellow marbles. If a single marble is chosen, what is the
probability that the marble is:
a. Red – 5/10 or 50%
b. Green – 3/10 or 30%
c. Blue – 2/10 or 20%

The probability of an event is a number between 0 and 1 where 0 represents zero chance of happening while 1
represent certainty of happening.

It is the product of the probability that an event will occur and the amount to be received upon such an occurrence.

Expected Value = Probability * Amount to Received in the occurrence of the event

Exercise on Expected Value

In planning its budget for the year, VV Company prepared the following payoff probability distribution describing
the relative likelihood of monthly sales volume levels and related contribution margin for product Violet
Monthly sales Volume Contribution Margin Probability
4,000 P 80,000 .20
6,000 120,000 .25
8,000 160,000 .30
10,000 200,000 .15
12,000 240,000 .10

What is the expected value of the monthly contribution margin for Product Violet?

5
Value of Perfect Information – The expected value of perfect information is the difference between the expected
value without perfect information and the return if the best action is taken given perfect information.

Expected value with perfect information

Less Expected Value of the best choice
Net Expected Value of Perfect Information

Payoff Tables – A payoff table illustrates all the possible scenarios presented together with all alternative courses
of action and probabilities of occurrence.

Illustration:
A dealer in luxury yachts may order 0, 1 or 2 yachts for this season’s inventory but no more or less. There is
$50,000 cost for carrying each excess yacht, and a $200,000 gain for each yacht sold. The situation may be
described by a payoff table as follows:

Actual Demand Order 0 Order 1 Order 2 Probability

0 yachts 0 $(50,000) $(100,000) 10%
1 yacht 0 200,000 150,000 50%
2 yachts 0 200,000 400,000 40%

The Expected Values of each course of actions are as follows:

Order 0 = $0 x 10% + $0 X 50% + $0 X 40% = $0

Order 1 = [$(50,000) x 10%] + [$200,000x50%] + [$200,000 x 40%] = $175,000
Order 2 = [$(100,000) x 10%] + [$150,000x50%] + [$400,000 x 40%] = $225,000

If the dealer had perfect knowledge of demand, he would make the best decision for each state of nature. The cost of
the other decision is the conditional cost of making other than the best choice. This cost may be calculated by
subtracting the expected value from the expected value given perfect information. This difference measures how
much better off the decision maker would be with perfect information.

Expected value with Perfect Information

($0 x 10% + $200,000x 50% + $400,000 x 40%) $260,000
Expected Value of the best Choice (225,000)
Expected Value of Perfect Information $ 35,000
========

The dealer will not pay more than $35,000 for information about future demand because it would then be more
profitable to make the expected value choice than to pay more for information.

Exercise:
A beverage stand can sell either soft drinks or coffee on any given day. If the stand sells soft drinks and the weather
is hot, it will make P2,500; if the weather is cold, the profit will be P1,000. If the stand sells coffee and the weather
is hot, it will make P1,900; if the weather is cold, the profit will be P2,000. The probability of cold weather on a
given day at this time is 60%

Required: Set up a pay off table and determine the expected values of selling softdrinks and coffee. Determine the
expected value of perfect information.

6
 4. Network Analysis

It is a range of techniques in the planning and management of projects wherein time is the primary consideration. It
is most useful under the following cases:

a. Time of project completion is an important consideration.

b. When the project is composed of different stages.
c. When time required for completion of the component stages is not known with certainty.
d. When resources used in the project completion.

Applications:
a. Physical distribution systems
b. Communication networks
c. City streets and traffic signals
d. Airline flight legs

Node – represents either the starting or destination/ terminal point in a network

Arc – usually represents distance but can also represent times or costs.

Event - represents a specified accomplishment at a particular instant in time.

Activity - represents the time and resources necessary to move from one event to another.

Major Variation of the Network Analysis

a. (Henry L. ) Gantt Chart or bar chart - dividing the project into logical subprojects called activities or tasks and
estimating the start and completion times for each activity and depicting them using horizontal bars along a time
scale.

Applications:
a. Construction
b. New product introduction
c. Systems installation
d. Personnel Staffing

Illustration of Gantt Chart

Activity Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 Week 7

A
B
C
D
E
F

Activity Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 Week 7

A 8,000.00 8,000.00
B 15,000.00
C 10,000.00 10,000.00 10,000.00
D 30,000.00 30,000.00
E 40,000.00 40,000.00
F 60,000.00 60,000.00
TOTAL 8,000.00 23,000.00 10,000.00 10,000.00 80,000.00 130,000.00 60,000.00

b. Program Evaluation and Review Technique (PERT)

This technique evolved from the use of Gantt charts in the late 1950s and was first applied to the US Navy’s Polaris
submarine project.

PERT - a probabilistic diagram of interrelationships of a complex series of activities; the major burden of which is
the determination of the critical path

7
 *Expected Time (te) = to + 4tm + tp to - optimistic time
--------------------- tm - most likely
6 tp - pessimistic time

*Expected Time - average time an activity would require if it were repeated a large number of time.

PERT- Critical Path Method (CPM) - used to estimate duration of project activities.

Critical path - defined as the longest path through the network.

Slack - length of time by which a particular activity can slip without having any delaying effect on the end
event.

Critical Path Method – developed independently of PERT and is widely used in the construction industry. Like
PERT, it is a network technique; but unlike PERT, it uses deterministic time and cost estimates. Its advantages
include cost estimates plus the concept of crash efforts and costs.

Crashing of activities  only activities under the critical path are to be crashed (reduce time) and only if the
reduced time will be cost effective.

Exercise on Network Analysis

Following are a set of activities for a project, their predecessor restrictions and three estimates of completion time.

Activity Predecessors Time estimates in weeks

Optimistic Most likely Pessimistic
A none 1 3 5
B none 1 2 3
C none 3 5 7
D A 3 4 5
E B 3 4 5
F B 5 6 13
G C 5 6 13
H D, E 2 8 14
I G, F 5 6 13

Required:
1. Prepare a PERT network
2. Identify the critical path

A D
H
E END
B
F
I
C
G

5. Decision Tree

It is defined as an analytical tool used in a problem in which a series of decisions has to be made at various time
intervals; with each decision influenced by the information that is available at the time it is made.

Decision points - represented by squares; the point where a decision has to be made

Chance events - represented by circles; presented as effects/results of the decision made at the decision point

8
Probabilities - presented as percentages of the outcome

Exercises on Decision Tree Analysis

*** The management of Sequoia Music is trying to decide whether to open a new store in SM Megamall, Tutuban
Center or Robinson's Place. Demand for the company's product in the new area is uncertain, but the marketing
manager has assigned probabilities to the three levels of demand.

These probabilities, as well as the contribution margin (conditional values in millions of pesos) for each store and
demand level are as follows:

Demand Level
Location High Moderate Low
SM Megamall P7 P2 P -1
Tutuban Center 6 3 0
Robinson's Place 5 4 4

Probability .30 .50 .20

Required: 1. Construct a decision tree for the above situation.

3. Identify the decision preference.

*** Trees Company is considering whether to develop and market a new product. Development costs are estimated
to be P100,000 and there is a 70% probability that the development effort will be successful. If the development is
successful, the product will be marketed and it is estimated that:

a. If the product is highly successful and the probability is 40%, it will produce a differential income of P400,000 (or
a net of P300,000 after subtracting the development cost)

b. If the product is moderately successful and the probability is 40%, it will break even, that is, its income of
P100,000 will just offset the development cost.

c. If the product is a failure and the probability is 20%, it will lose P200,000 after taking into account the
development cost.

Required: 1. Prepare a decision tree showing the alternatives.

2. Should the company develop the product or not?

9
 6. Queuing Theory

It is a group of mathematical models for systems involving waiting lines and is used to determine the operating
characteristics of a waiting line such as:
a. the average units in the line
b. average time a unit waits
c. the probability that a unit must wait
d. the average time a unit is in the system.

Eg. Bank teller windows, grocery checkout counters, highway toll booths.

The objective of the queuing theory is to minimize the total cost of the system including those associated with
providing service and waiting costs for a given rate of arrivals.

Service
Calling population facility

Queue Served calling units

System time

Other key terms:

1. Phases – number of steps/processes that a unit will have to pass through

2. Channels/ Servers– number of available lines to provide service

Configurations:
1. Single Phase – Single Channel  newspaper vendor
2. Single Phase – Multiple Channel  fastfood
3. Multiple Phase – Single Channel  drive through
4. Multiple Phase – Multiple Channel  NBI clearance request

7. Markov Chain

A powerful tool in business used to analyze the current status of a product in an effort to predict its future. It is a
useful technique in marketing which examines and forecasts the behavior of customers from the standpoint of their
loyalty to one department store and their habits of switching to other stores.

The outcome of any particular experiment depends only on the results of the immediately preceding experiment and
not on any other experiment conducted in the past.

Illustration:
1. Studies in a two-grocery store town have shown the following probabilities of customers buying from one
store one month and the other store the next. If a customer buys from A this month, there is a 60%
probability he will return next month. If he buys from B, there is an 80% probability of return. What is the
probability of buying from A in period 3, if bought from A this month? Predict the future state from the
current state and the transition state.

Now Next Month Third Month

(Current) (Transition)
60%A 60%A (60% X 60%) = 36%

A 100% 40%B

40%B 20% A (40%X20%) = 8%

10
 80% B ------
44%

Probability of purchasing from A in period 3 if bought from A this month = 44%

2. The copy machine in an office is very unreliable. If it was working yesterday, there is an 80% chance it
will work today. If it was not working yesterday, there is a 10% chance it will work today.

a. What is the probability that it is not working today, if it was not working yesterday?

b. What is the probability it will not work today, if it was working yesterday?

c. If it is working today, what is the probability that it will be working 2 days from now?

d. If it is not working today, what is the probability that it will be working 2 days from now?

8. Game Theory

It is a mathematical approach to decision making when confronted with an enemy or competitor.

Decision rules developed by game theorists are:

a. Maximax Criterion – adopted by risk seeking optimistic players who desire the largest possible payoff and
are willing to accept high risk. (Maximum of the Maximum Payoffs)

Conditional Profit (Payoff) Table

States of Nature Actions: Go Into New Market

Alabang Bacoor Cavite Divisoria
High 18 9 14 12
Medium 12 6 10 11
Low 3 5 6 9

The Alternative to be chosen is: ___________

b. Maximin Criterion – determines the maximum loss for each decision possibility and then chooses the
decision with the minimum maximum loss; used by risk averse players for whom the utility of a gain is less than the
disutility of an equal loss. (Maximum of the Minimum Payoffs)

Conditional Profit (Payoff) Table

States of Nature Actions: Go Into New Market

Alabang Bacoor Cavite Divisoria
High 18 9 14 12
Medium 12 6 10 11
Low 3 5 6 9

The Alternative to be chosen is: ___________

c. LaPlace (Average Profit) criterion – Calculate the average profit fro each alternative action. Then choose
the action having the highest average profit. All states of nature are thus treated as equally likely to occur.

Conditional Profit (Payoff) Table

11
States of Nature Actions: Go Into New Market
Alabang Bacoor Cavite Divisoria
High 18 9 14 12
Medium 12 6 10 11
Low 3 5 6 9

Average 11 6.67 10 10.67

The Alternative to be chosen is: ___________

d. Minimax Regret criterion – player who wishes to minimize the effect of a bad decision in either direction
chooses the decision that has the lowest maximum opportunity cost.

Procedure: Within each row, subtract the profit for each alternative from the maximum profit in that row. Find the
maximum regret for each alternative action. Then select the alternative with the smallest maximum regret.

Regret (Opportunity Cost) Table

States of Nature Actions: Go Into New Market

Alabang Bacoor Cavite Divisoria
High 0 9 4 6
Medium 0 6 2 1
Low 6 4 3 0

The alternative to be chosen under this criteria is: ___________

12