Chapter 06 – Decision Model

Learning Objectives
After completing this chapter, you will be able to:

• Know the steps of the decision-making process.

• Describe the types of decision-making environments.
• Make decisions under uncertainty.
• Use probability values to make decisions under risk.
• Revise probabilities using Bayesian analysis.
• Develop accurate and useful decision trees.
• Use POM-QM to solve basic decision-making problems.
Decision Analysis – The Concept

 A decision is a choice among alternatives based on estimates of

the values of those alternatives.

Good Decision Bad Decision

• based on reasoning • Not based on reasoning
• consider all available data and • Do not consider all available data
possible alternatives and possible alternatives
• employ a quantitative approach • Do not employ a quantitative
approach
• occasionally result in an • occasionally result in a good
unexpected outcome; it is still a outcome if you are lucky; it is still a
good decision if made properly bad decision

Decision Theory - an analytic and systematic approach to decision

making.
Decision Analysis – The Concept

 Decision analysis is a method used to develop an optimal

strategy when a decision maker is faced with several
decision alternatives with an uncertain or risk-filled pattern of
future events.
 A decision problem is characterized by:
– Decision alternatives: a course of action or strategy that may be
chosen by the decision maker ,
– States of nature: an outcome over which the decision maker
has little or no control , and
– Resulting payoffs (cost or revenues): A reward for all possible
combinations of alternatives and states of nature
`
The goal of decision analysis is to make a choice among alternatives
that optimizes the resulting payoff based on a decision criterion.
Decision Analysis – The Concept

The six steps in Decision Analysis:

1. Clearly define the problem (to maximize revenue or
minimize cost?)

2. List the possible alternatives (actions/decisions)

3. Identify the possible outcomes (state of nature)

4. List the payoff (profit/reward or cost)

5. Select one of the decision theory models (on the basis of

the operating environment and degree of uncertainty).

6. Apply the model and make your decision

Decision Analysis – Payoff Table Analysis

 Decision analysis can be made by:

 Payoff Table
 Decision Tree
Decision Analysis – Payoff Table Analysis
 Payoff table analysis can be applied when:
 There is a finite set of discrete decision alternatives.
 The outcome of a decision is a function of a single future event.
 In a Payoff table -
 The rows correspond to the possible decision alternatives.
 The columns correspond to the possible future events.
 Events (states of nature) are mutually exclusive and collectively
exhaustive.
 The table entries are the payoffs.
Decision Analysis – Payoff Table Analysis

 Example: An investor would like to invest on four potential

projects.
– Project A
– Project B
– Project C
– Project D
 The return on each investment depends on the (uncertain)
market behavior during the year
 The investor builds a payoff table to help make the
investment decision.
Decision Analysis – Payoff Table Analysis

 The payoff table shows potential “payoff” depending upon

likely economic conditions.

Payoff Table
Decision Economic Condition/State of Nature
Alternatives Recession Normal Boom
Project A 4075 5000 6100
Project B 0 5250 12080
Project C 2500 7000 10375
Project D 1500 6000 9500

 Note: Since the payoff in project C is higher than the payoff for D for
every economic condition, project C is dominant. We can eliminate
project D from consideration.
Decision Making Environments

 Decision-making environments are classified into:

1. Decision making under certainty.
 The future state-of-nature is assumed known.

2. Decision making under uncertainty.

 There is no knowledge about the probability of the
states of nature occurring.
 Various outcomes are possible
3. Decision making under risk.
 There is some knowledge of the probability of the
states of nature occurring.
1. Decision Making under certainty

 Here, the state of nature is certain (one state).

 Only one outcome for each decision alternative.
 Select decision that yields the highest return

All the topics we have covered so far!

Optimization Problems
2. Decision Making under Uncertainty

 Here, state of nature is uncertain (several possible states).

 Various outcomes are possible for each decision alternative.
 Decision maker cannot assign probabilities to the
States of Nature.
 Many business problems contain uncertain elements that are
impossible to ignore.
Example: Construction of different model houses for sale
(Not certain about the market demands).

Decision is made based on chosen decision criteria.

2. …under Uncertainty

 The decision criteria are based on the decision maker’s attitude

towards life (Optimistic, pessimistic, neither both/realists).

A Glass of Water:
Optimist: The glass is half full.
Pessimist: The glass is half-empty.
Realist: The glass is twice as large as it needs to be.

All are right!

 The optimist sees opportunity in every problem.

 The pessimist sees problem in every opportunity.

 The realist uses the trade offs!

All contribute to our society! The optimists invent the airplane and the
pessimists the parachute; the realists improve it!
2. …under Uncertainty

 The decision criteria include:

 Maximax Criterion - seeks the largest of maximum payoffs.
 Maximin Criterion - seeks the largest of the minimum payoffs
among the actions.
 Minimax Regret Criterion - seeks the smallest of the maximum
regrets among the actions.
 The Criterion of Realism – seeks a weighted average where
maximum and minimum rewards are weighted by some coefficient.
 Principle of Insufficient Reasoning – seeks the largest payoffs
among the sum obtained on each alternative across all events.

Note: But, the outcome is still uncertain. Uncertain parameters become

known only after a decision is made.
i) The Maximax Criterion

 This criterion is based on the best possible scenario.

 It fits both an optimistic (Risk Taking) and an aggressive
decision maker.
 An optimistic decision maker believes that the best possible
outcome will always take place regardless of the decision made.
 An aggressive decision maker looks for the decision with the
highest payoff (when payoff is profit).
 To find an optimal decision.
 Find the maximum payoff for each decision alternative.
 Select the decision alternative that has the maximum of
the “maximum” payoff.
 i) The Maximax Criterion …

If you are an optimist, you will decide on the basis of Maximax.

Alternatives Economic Condition Step 1: Pick the max value

Recession Normal Boom for each alternative.
Project A 4075 5000 6100 6100
Project B 0 5250 12080 12080
Project C 2500 7000 10375 10375

Step 2:Then pick the alternative

with max payoff.
ii) Maxmin criterion

 This criterion is based on the worst-case scenario.

 It fits both a pessimistic (Risk Averse) and a
conservative decision maker’s styles.
o A pessimistic decision - the worst possible result will
always occur.
o A conservative decision - ensure a guaranteed minimum
possible payoff.
ii) Maxmin criterion…

 To find an optimal decision

 Record the minimum payoff across all states of nature for each
decision.
 Identify the decision with the maximum “minimum payoff.”

Alternatives Economic Condition 1: Pick the min value

Recession Normal Boom for each alternative.
Project A 4075 5000 6100 4075
Project B 0 5250 12080 0
Project C 2500 7000 10375 2500

2: Then pick the alternative

with max payoff.
iii) The Minimax Regret Criterion

 The Minimax Regret Criterion again fits both a pessimistic

and a conservative decision maker approach.
 The payoff table is based on “lost opportunity,” or “regret.”
 The decision maker incurs regret by failing to choose the
“best” decision.
If I knew the future, how much I’d regret
for my decision…
iii) The Minimax Regret Criterion …

 To find an optimal decision, for each state of nature:

 Determine the best payoff over all decisions.
 Calculate the regret as the difference between its payoff
value and this best payoff value.
 For each decision find the maximum regret over all states
of nature.
 Select the decision alternative that has the minimum of
these “maximum regrets.”
iii) The Minimax Regret Criterion …

Alternatives Economic Condition Step 1: Calculate

the maximum for
Recession Normal Boom each outcome.
Project A 4075 5000 6100
Project B 0 5250 12080
Project C 2500 7000 10375
4075| 7000| 12080

Alternatives Regret Table

(opportunity loss Table) Stet 2: Prepare
Recession Normal Boom “Regret Table” by
Project A subtracting each
0 2000 5980
outcome cell value
Project B 4075 1750 0 from its maximum.
Project C 1575 0 1705
iii) The Minimax Regret Criterion …

Alternatives Regret Table Step 3: Pick the max

Recession Normal Boom value for each alternative.
Project A 0 2000 5980 5980
Project B 4075 1750 0 4075
Project C 1575 0 1705 1705
Step 4: Pick the alternative
with minimum regret.
v) The Criterion of Realism

 Also known as the weighted average or Hurwicz criterion.

 A compromise between an optimistic and pessimistic decision (Risk

Tradeoff) .

 A coefficient of realism,  (0 <  <1), is selected by the decision

maker to indicate optimism or pessimism about the future
– When  is close to 1, the decision maker is optimistic.
– When  is close to 0, the decision maker is pessimistic.

Criterion of realism = (row maximum) + (1- )(row minimum)

 A weighted average where maximum and minimum rewards are

weighted by  and (1 - ), respectively.
v) The Criterion of Realism

Alternatives Economic Condition Step 1: (row maximum) +

Recession Normal Boom (1- )(row minimum)
Project A 4075 5000 6100 5695
Project B 0 5250 12080 9664
Project C 2500 7000 10375 8800

Step 2:Then pick the alternative

with the highest weighted value.

Assume coefficient of realism of =0.80. Then, weighted Averages:

– Project A = (0.8)(6100) + (0.2)(4075) = 5695
– Project B = (0.8)(12080) + (0.2)(0) =9664
– Project C = (0.8)(10375) + (0.2)(2500) = 8800

 Decision: Select the alternative with the highest weighted value, i.e.
Project B
vi) The Principle of Insufficient Reason(Equal likelihood)

 This criterion might appeal to a decision maker who is neither

pessimistic nor optimistic.
 It assumes all the states of nature are equally likely to occur. i.e.
no information about their likelihood.
 The procedure to find an optimal decision.
 For each decision add all the payoffs.
 Select the decision with the largest sum (for profits).
Alternatives Economic Condition Step 1: Sum the payoffs for each
Recession Normal Boom alternative.

Project A 15175
4075 5000 6100
Project B 0 5250 12080 17330
Project C 2500 7000 10375 19875

Step 2:Then pick the alternative

with max payoff.
Sample Problem

1. A company would like to expand its production by setting up new

plants. The course of actions to be decided and the state of nature
are shown in the following pay off table.

Alternatives Economic Condition

High Demand Moderate Demand Low Demand
Large Plant 200,000 100,000 -120,000
Small Plant 90,000 50,000 -20,000
No Plant 0 0 0

Make a decision based on different decision criteria.

Let us do it with excel

3. Decision Making under Risk – i) The expected
value (EV) criterion
 The above four approaches we used involved Decision
Making without Probabilities.
 The probability estimate for the occurrence of
each state of nature (if available) can be incorporated in the
search for the optimal decision.
 For each decision, calculate the expected payoff as follows:

Expected Payoff = (Probability)(Payoff)

(The summation is calculated across all the states of nature)

 Select the decision with the best expected payoff.
 It is a kind of weighted average method.
i) The Expected value (EV) Criterion…

Decision State of Nature Expected

Alternatives Recession Normal Boom Value
Project A 4075 5000 6100 5090
Project B 0 5250 12080 5907.50
Project C 2500 7000 10375 6943.75
Prior Prob. 0.20 0.55 0.25
EV = (0.2)(4075) + (0.55)(5000) + (0.25)(6100) = 5090

EV = (0.2)(0) + (0.55)(5250) + (0.25)(12080) = 5907.50

EV = (0.2)(2500) + (0.55)(7000) + (0.25)(10375) = 6943.75

i) The Expected value (EV) Criterion…
 The expected value criterion is useful generally in two cases:
 Long run planning is appropriate, and decision situations
repeat themselves.
 The decision maker is risk neutral.
ii) The Expected Value of Perfect Information (EVPI)

 Frequently information is available which can improve the

probability estimates for the states of nature.
 May be interested to purchase it (or, hire consultant) and thus make
a better decision
How valuable would this information be?
 EVPI is the gain in expected return obtained from knowing with
certainty the future state of nature.
EVPI  ERPI  EREV
where,
ERPI  exp ected return with perfect inf ormation
EREV  exp ected return by EV criterion
 The EVPI provides an upper bound on the expected value for
additional information. i.e. The maximum amount you would pay to
the source/consultant to provide information.
ii) The EVPI…

 If it was known with certainty that there will be a “Boom” in the

economic condition, …

Decision Economic Condition

Alternatives Recession Normal Boom
Project A 4075 5000 6100
Project B 0 5250 12080
Project C 2500 7000 10375
Prior Prob. 0.20 0.55 0.25

B
... the optimal decision would be to invest in ….

Similarly,…
ii) The EVPI…

Decision State of Nature Expected

Alternatives Recession Normal Boom Value

Project A 4075 5000 6100 5090

Project B 0 5250 12080 5907.50
Project C 2500 7000 10375 6943.75
Prior Prob. 0.20 0.55 0.25
Expected Return with Perfect information (ERPI)
= Σ(best payoff in state of nature i) (probability of i)
= 0.2(4075)+0.55(7000)+0.25(12080) = 7685
Expected return without additional information= EV criterion = 6943.75
EVPI = ERPI - EREV = 7685 – 6943.75 = 741.25

Note:
 “Perfect” information is extremely rare.
 You would be willing to pay some amount less than $741 depending on
how reliable the information is perceived to be.
Sample Problem

1. A company would like to expand its production by setting up new

plants. The course of actions to be decided and the state of nature
are shown in the following pay off table.
Alternatives Economic Condition
High Demand Moderate Demand Low Demand
Large Plant 200,000 100,000 -120,000
Small Plant 90,000 50,000 -20,000
No Plant 0 0 0
Probability 0.30 0.50 0.20

• Make a decision based on EV criterion.

• What is EVPI?

Let us do it with excel

Decision Trees
 The Payoff Table approach is useful for a non-sequential or
single stage.
 However, many real-world decision problems consists of a
sequence of dependent decisions.
 Decision Trees are useful in analyzing multi-stage decision
processes.
 A Decision Tree is a chronological representation of the
decision process.
Decision Trees - Structure
 The Decision Tree is composed of nodes and branches.

A branch emanating from a decision

Chance node corresponds to a decision
node
alternative is Decision Branch (DB). It
P(S2) includes a cost or benefit value.
Decision
node

A branch emanating from chance

node corresponds to a particular state
P(S2)
of nature, and includes the probability
of this state of nature is Chance
Branch (CB).

- A decision node is represented by a rectangle

- A chance node is represented by a circle.
- At the end of each branch, there is an end node represented by a triangle.
Decision Trees - Structure
 A decision tree is constructed from left to right.
 The decision tree must show all the possible paths that the
decision maker might follow through time. Including all
possible decision alternatives.
Decision Trees - Analysis

Six Steps of Decision Tree Analysis:

1. Define the problem
2. Structure/draw the decision tree
3. Assign probabilities and payoff to the states of nature (chance
branch)
4. Compute the expected payoff at each end node.
5. At the chance node, calculate the average (i.e. expected)
payoff, which is usually termed as Expected Monetary Value
(EMV). If there is no chance event for a particular decision
branch, it’s EMV is equal to the payoff.
6. Select the decision with the highest EMV for each chance
node. Otherwise, if dealing with costs, select the decision with
the lowest EMV.
Decision Trees - Example

 Suppose you bought 500 units of X at $10 each. A dealer

has offered to buy these from you at $14 each (you can
make $4/unit profit).
 Alternatively, you can sell these yourself for $16 each
($6/unit profit) but the demand is uncertain. The demand
distribution is shown in the table.

Demand: X 300 400 500 600

Pr(X) 0.30 0.45 0.20 0.05

 Note: if demand exceeds 500, you will sell all 500. On the other
hand, if demand is under 500, you will have leftover units. These
leftover items can disposed off for $7 each ($3 loss, the dealer will
no longer buy these leftover units from you).

What’s your decision?

Decision Trees - Solution

 Start with the tree having 2 branches (DB) at the decision

point.
 There are no chance events in the dealer sale branch,
 For the self sale, there are 4 mutually exclusive possibilities
(state of nature).

Dealer
Sale

Self sale
500, 20%
Decision Trees - Solution

Payoff = 500*4 = 2000 EMV = 2000

Dealer
Sale
Payoff = 300*6 – 200*3 = 1200

Payoff = 400*6 – 100*3 = 2100

Self sale
500, 20%
Payoff = 500*6 = 3000

Payoff = 500*6 = 3000

EMV = 0.3*1200 + 0.45*2100 +

0.2* 3000 + 0.05*3000 Your decision? Self sale
= 2055
THANK YOU
Assignment

1. What are the elements for rational decision making?

2. Why decisions are hard?
3. What is a good decision? A bad decision?
4. A broker with $1000 has five potential investments A, B, C, D and
E. The return on each investment depends on the (uncertain)
market behavior during the year. The broker built a payoff table to
help make the investment decision. Find an optimal decision
based on:
a) Maximax Criterion;
b) Maximin Criterion
c) Minimax Regret Criterion; and
d) principle of Insufficient Reasoning
Assignment

Decision States of Nature

Alternatives Large Rise Small Rise No Change Small Fall Large Fall
A -100 100 200 300 0
B 250 200 150 -100 -150
C 500 250 100 -200 -600
D 60 60 60 60 60
E 200 150 150 -200 -150
Assignment

5. A building construction company is planning to construct

luxury condominiums on its site. The company has a
capacity to construct up to 250 condominium blocks. The
financial success of the project depends upon the number of
condominium blocks to be constructed and the chance
event concerning the demand for the condominiums. The
decision problem is to select the number of condominium
blocks to be constructed that will lead to the largest profit
given the uncertainty concerning the demand for the
condominiums. The cost and revenue of construction per
block are 6000 and 14000 (x1000)birr, respectively.
Assignment

Hint: Payoff Table

State of Nature

Alternative Demand
Actions Low (50 units) Medium (100 units) High (150 units)

Build 50 400,000 400,000 400,000

Build 100 100,000 800,000 800,000

Build 150 (200,000) 500,000 1,200,000

Find an optimal decision:

a) Maxmin Criterion
b) Minimax Regret Criterion
c) Maximax Criterion
d) Principle of Insufficient Reasoning