TOPIC 5: DECISION THEORY

5.1 INTRODUCTION
Life can be summarized as a product of the sum total of decisions made. These
choices are generally made in the midst of various alternatives such as choice for a
school, college, or university; choice for subjects, a course or a program; choice for a
career; choice of means production or technology; choice for a business; choice for
investment; choice for friends, spouse etc.

The performance of person or an individual entity like a household, firm, a country or

any other organization is a function of the quality of the decisions made. The quality
of decisions heavily depends on the decision making process. All managerial
activities involve the decision making process which entails the best course of action
amongst alternatives.

Decision theory may be defined as:

• A process which results in the selection, from a set of alternative courses of
action, that course of action which is considered to meet the objectives of the
decision problem more satisfactorily than others as judged by the decision
maker.
• The process of logical and quantitative analysis of various alternative courses
of action in order to arrive at the best strategy based on certain decision
criteria.

It is generally argued that objectivity rather than subjectivity, reason rather than
emotion, open adherence to certain verifiable concepts, principles and decision rules
rather than expedient opportunism, orderly precise processes rather than haphazard,
rough and ready rule of thumb process, are the basic ingredients of decision-making
process. The method permits the decision maker to generate accurate empirical data,
to establish logical relationships among the various variables in the decision situation
and to make optimal decisions based on facts.

5.2 STEPS IN DECISION THEORY

A scientific decision-making approach consists of the following steps:

1. Identification and definition of the problem at hand

2. Specification of objectives and decision criteria.
3. Identification and evaluation of the possible alternatives
4. Formulation or selection of a mathematical decision theory model
5. Choice of the best alternative course of action through application of the
model.
6. Sensitivity analysis of the ‘solution’
7. Communication and implementation of decision.
8. Follow-up and feedback of results of decision.

1
5.3 CONCEPTS IN DECISION MAKING

1. Decision maker – individual or group of individuals responsible for making the

choice of an appropriate course of action.
2. Courses of action – alternative strategies available to the decision maker.
3. States of nature/outcomes – these are events whose occurrences determine the
level of success of a course of action but are beyond the control of a decision
maker.
4. Pay-off – the net benefit to the decision maker that accrues from a given
combination of decision alternatives and events.
5. Pay-off table – table comprising of the states of nature (outcomes or events) and a
set of given courses of action or strategies.

Illustration of pay-off table

Assume that we have:
m no. of events under states of nature S1, S2, …Sn
n alternative courses of action (strategies) a1,a2, …,an
Pay off for activity Aij under the state of nature Si is denoted by:
Pij (i=1, 2, …,m, j=1, 2, …n)

m  n pay-off table
Courses of Action (Strategies)
States of nature
A1 A2 ……………… An
S1 P11 P12 …………….. P1n
S2 P21 P22 ……………... P2n

Sm Pm1 Pm2 ……………… Pmn

Weighted profit is the product of the pay-off of a state of action and the probability of
occurrence of the given state of nature.
weighted profit = Pij P( S i )
5.4 Regret or opportunity loss table – it is the difference between the highest
possible profit for a state of nature and the actual profit obtained for the particular
action or strategy taken i.e. the loss incurred for not choosing the best possible
course of action or strategy.

Assume m is the best strategy or optimum value.

Courses of Action (Strategies)

States of nature Conditioned opportunity loss
A1 A2 ……………… An
S1 M1 – P11 M1 – P12 …………….. M1 –P1n
S2 M2 – P21 M2 –P22 ……………... M2 –P2n

Sm Mm – Pm1 Mm –Pm2 ………………. Mm –Pmn

2
Example 1
A trader buys and sells a certain product in the range of 0 to 4 units (0, 1, 2, 3, 4). He
buys each unit at ₤ 40 and sells it at ₤ 45. The demand could be between 0 and 4 units.
Generate the following:
i) A conditional pay-off table
ii) The regret (opportunity loss) table

A conditional pay-off table

Courses of Action
State of nature
Conditioned pay-offs/ Possible supply
(Probable demand)
0 1 2 3 4
0 0 -40 -80 -120 -160
1 0 5 -35 -75 -115
2 0 5 10 -30 -70
3 0 5 10 15 -25
4 0 5 10 15 20

The regret (opportunity loss) table

Courses of Action
State of nature
Conditional opportunity loss
(Probable demand)
0 1 2 3 4
0 0 40 80 120 160
1 5 0 40 80 120
2 10 5 0 40 80
3 15 10 5 0 40
4 20 15 10 5 0

5.5 TYPES OF DECISION MAKING ENVIRONMENTS

The main aim of decision theory is to help the decision maker in selecting the best
course of action from amongst available strategies. The decision models can be
classified into 4 types: certainty, risk, uncertainty and conflict
i) Decision making under certainty
There is complete and accurate knowledge of the outcome of each alternative.
There is only one outcome for each alternative.
ii) Decision making under risk
There are multiple possible outcomes of each alternative and a probability of
occurrence can be attached to each.
iii) Decision making under uncertainty
There are multiple possible outcomes of each alternative but there is no
knowledge of the probability of occurrence that can be attached to each.

iv) Decision making under conflict

I) Decision making under certainty (deterministic model)

The decision maker knows with certainty the consequences of selecting each
course of action. The data is well defined and the decision maker is very sure on

3
 what to base the decision on. The techniques used in this case include break-even
analysis, inventory models under certainty, linear programming, transportation,
and assignment models.

II) Decision making under risk (probabilistic model)

In this case the payoffs associated with each decision alternative are described by
probability decisions. The decision making is based on expected value criterion in
which the alternatives are compared based on maximization of expected gain or
minimization of expected loss.

Expected value (EV)

The expected value (EV) of an event is the product of its probability of occurrence
of the event and the outcome, or the value of the event (or a series or trial).
Example 2: a course of action has 0.3 chance of occurrence resulting in a gain of
shs.1200. What is the expected value?
EV = probability  outcome
= 0.3  1200
= 360
Expected monetary value (EMV)
It is the sum of the product of each possible outcome of a course of action with its
probability of occurrence.

Suppose on the basis of prior knowledge due to either past experience or on a

subject basis, that the state of nature Si has the probability of occurrence P(Sj)
[j=1, 2, …, k] then the expected monetary value corresponding to the courses of
action of the decision maker is given by:
EMV ( Ai ) = pi1 P( S1 ) + pi 2 P( S 2 ) + pi 3 P( S 3 ) + ... + pij P( Sj )
Steps for calculating EMV
1. Construction of a pay-off table listing the various courses of action (Ai) and states
of nature (Si).
2. List the pay-off associated with each possible combination of course of action and
the states of nature along with the corresponding probability of each state of
nature.
3. Calculate EMV for each course of action by multiplying weighted conditional
profits/losses (pay-offs) by the associated probabilities and sum this weighted
value for each course of action.
4. Determine the course of action which corresponds to the EMV on the basis of a
specified decision.

Note
• EMV criterion is also referred to as the Bayes’ decision rule. It is frequently
used in decision-making to find the best strategy.

Example 3
Using the example 1 and the probabilities given below find the best course of action
for the trader on the basis of EMV using
a) Expected payoffs.
b) Expected opportunity losses (regrets)

4
 a) Expected payoffs
Computation of expected monetary value

Conditional pay-off Expected pay-offs

State Course of action Course of action
Probability
of
P(Sj) A1 A2 A3 A4 A5 A1 A2 A3 A4 A5
nature
(0) (1) (2) (3) (4) (0) (1) (2) (3) (4)

S1(0) 0.04 0 -40 -80 -120 -160 0 -1.6 -3.2 -4.8 -6.4

S2(1) 0.06 0 5 -35 -75 -115 0 0.3 -2.1 -4.5 -6.9

S3(2) 0.20 0 5 10 -30 -70 0 1 2 -6 14

S4(3) 0.30 0 5 10 15 -25 0 1.5 3 4.5 -7.5

S5(4) 0.40 0 5 10 15 20 0 2 4 6 8

EMV 0 3.2 3.7 -4.8 -26.8

The maximum value of EMV corresponds to the course of action A3. Hence on the
basis of EMV criterion the dealer can opt for strategy A3 i.e. buy 2 units of the
product which yield the maximum EMV, ₤ 3.7.

Example 4
The cost of making an item in a certain industry is ₤ 25. The selling price of the item
is ₤ 30, if it is sold within a week and could be disposed off at ₤ 20 per item at the end
of the week. The sales for a certain period were recorded as follows:
Weekly
3 4 5 6 7 8
sales
No. of
0 10 20 40 30 0
weeks

Find the optimum number of items per week that the industry should produce.
Cost of production = ₤ 25
Selling price = ₤ 30
Selling price at the end of the week = ₤ 20
Favourable chances
P( S j ) =
Sum of possible outcomes
Profit = ₤ 5 per item (30-25)
Loss = ₤ 5 per item (20-25)

5
 State of
Nature Probability conditional pay-off expected pay-off
(Demand) P(Sj) (production per week) (production per week)
(Sj)

4 5 6 7 4 5 6 7

S1(4) 0.1 20 15 10 5 2 1.5 1 0.5

S2(5) 0.2 20 25 20 15 4 5 4 3

S3(6) 0.4 20 25 30 25 8 10 12 10

S4(7) 0.3 20 25 30 35 6 7.5 9 10.5

EMV 20 24 26 24

The best course of action is to produce 6 items because of the high EMV (₤ 26).

Exercise

1. The payoffs of three strategies A1, A2 and A3 and the possible states of nature
S1, S2 and S3 are given in the table below. The probabilities of the states of
nature are 0.3, 0.4 and 0.3 respectively. Determine the optimal strategy using
the expectation principle.

States of Nature A1 A2 A3
S1 -20 -50 200
S2 200 -100 -50
S3 400 600 300

2. A vendor buys weekly newspapers at kshs. 30 each and sells it at kshs. 50

within the week and kshs. 30 after the week. The data on the past sales are as
follows:

No. of newspaper sold within the 400 500 600 700

week
No. of weeks 10 20 40 30

Find the optimum number of newspapers that the vendor should buy every
week in order to maximize the profit.

6
3. The management of a company is faced with a problem of choosing any of the
3 products for manufacturing. The potential demand for each product would
turn out to be good, moderate or poor. The probability of each of the states of
nature was estimated as follows.

Product Good Moderate Poor

X 0.70 0.20 0.10
Y 0.50 0.30 0.20
Z 0.40 0.50 0.10
The estimated profit or loss under the 3 states is:
Product Good Moderate Poor
X 30,000 20,000 10,000
Y 60,000 30,000 20,000
Z 40,000 10,000 -15,000
Required: prepare the EV table and advise the management about the choice
of the product.

III) DECISION MAKING UNDER UNCERTAINTY

This involves alternative courses of action whose pay-offs depend on the
random state of nature whose probability of occurrence is either unknown or
cannot be determined. Such situations may arise when a new product is
introduced in the market. The choice of a course of action heavily depends on
the personality of the decision maker and the policy of an organization.
The decision criteria used under uncertainty include the following:

1. Maximin or pessimism (Waldian decision-making criteria)

2. Maximax or optimism criterion
3. Minimax regret criterion
4. Hurwic or realism criterion
5. Laplace or Baye’s or rationality criterion

1. Maximin (Waldian decision-making criteria)

It is the decision to take the course of action that which maximizes the minimum
possible payoffs i.e.the best of the worst. It is a decision rule based on maximizing
the minimum returns that occur.
Procedure:
▪ Determine the minimum outcomes of every course of action
▪ choose the strategy with the maximum number among the minimum

courses of action
State of nature
A1 A2 A3 A4 A5
Sj
0 1 2 3 4
0 0 -40 -80 -120 -160
1 0 5 -35 -75 -115
2 0 5 10 -30 -75
3 0 5 10 15 -25
4 0 5 10 15 20
Minimum in columns 0 -40 -80 -120 -160

7
 Answer: Strategy A1

2. Maximax criterion (optimism)

It is the decision to take the course of action that which maximizes the maximum
possible payoffs i.e. best of the best. A decision rule based on maximizing the
maximum returns that can be gained
Procedure:
▪ list down the maximum or best payoffs for each course of action
▪ choose the maximum of the maximum values

A1 A2 A3 A4 A5
maximum payoffs 0 5 10 15 20
Answer: Strategy A5

3. Minimax/regret criterion (pessimism)

It is a decision rule that seeks to minimize the maximum regret that could be from
choosing a particular strategy.
regret payoff = max imum payoff − payoff
= M ij − Pij
Procedure:
▪ obtain the maximum regret of each course of action
▪ select the course of action with the minimum of the maximum regret values
A1 A2 A3 A4 A5
maximum regret 20 40 80 120 160

Answer: Strategy A1

4. Hurwic criterion
It is the weighted average of the best and worst payoffs of each action. It is a
compromise between the optimistic and pessimistic decision criteria.
Procedure:
• select the coefficient of optimism, 
 (0    1) when  is close to 1 the decision maker is optimistic about the
future, and when near 0 he is pessimistic.
• Select the strategy which maximizes:
H =  (max imum payoff in column) − (1 −  ) min imum payoff in the column

5. Laplace or Baye’s criterion (criterion of rationality)

Decision rule is based on the maximum average for every course of action. This is
a compromise between optimistic and pessimistic decision criterion.
Procedure:
▪ Calculate the average outcome for every course of action
▪ Select the maximum.

A1 A2 A3 A4 A5
Average 0 -4 -17 -39 -71

Answer: Choose A1 because it is the best of the averages.

8
 ASSIGNMENT 5

Q1. A business lady intends to invest all her funds in one of three alternative
investment plans namely stock, bonds and debentures whose payoff matrix
based on three potential economic conditions is as shown below:

Alternative
Boom Normal Recession
investment
Stock 12,000 9,000 6,000
Bonds 7,200 9,600 1,200
Debentures 6, 900 9,300 5,400

Determine the best investment plan using each of the following criteria:

a) Maximax

b) Maxmin

c) Laplace

Q2. A hawker buys a unit of her merchandise at $ 10 and sells it at $ 15 during the
day and $ 8 the following day. The distribution of the daily demands based
on past data is as shown below:

Daily 7 8 9 10 11
Demand
Probability 0.20 0.20 0.25 0.15 0.20

Determine her optimal strategy based on the expected monetary value

criterion.
Q3. The daily demand for a certain type of a cake at a retail shop are given by the
following probability distribution.

Demand 100 150 200 250 300

Probability 0.20 0.25 0.30 0.15 0.10

The price of a fresh cake is Kshs. 25. If a cake is not sold the same day, it is
certainly disposed of at Kshs. 10 at the end of the day. The cost per cake to
the shop is Kshs. 15. Assuming stock levels are restricted to one of the
demand levels, determine the recommended number of cakes that should be
stocked daily on the basis of the following decision criteria?
i) Maximin
ii) Maximax
iii) Laplace
iv) Expected monetary value