Ch. 1 Decision Theory

UNIT 1
DECISION THEORY &

DECISION TREES
Prof. Swapnil Udgir
SYLLABUS
1. Decision Theory & Decision Trees (7.5 Hours)
 Introduction to Decision-Making Process (0.5 Hours)
 Steps of Decision-Making Process (0.5 Hours)
 Types of Decision Making Environments (0.5 Hours)
 Decision Making under Uncertainty (1.5 Hours)
 Decision Making under Risk (3 Hours)
 Decision-Tree Analysis (1.5 Hours)

INTRODUCTION
 The success or failure that an individual or
organization experiences, depends to a large extent,
on the ability of making decisions on time.
 To arrive at such a decision,

the decision maker needs
to enumerate feasible and
viable courses of action
(alternatives or strategies),
the projection of conse-
quences associated with
each course of action.
INTRODUCTION
He also needs to find a measure of
effectiveness to identify the best course of
action.
A decision may be defined as the

selection of an act by the decision maker,
considered to be the best according to
some pre designated standard, from
amongst the available options.
INTRODUCTION
The decision theory is used to determine
optimal strategies where a decision maker is
faced with several decision alternatives and
an uncertain or risky pattern of future events.
COMPONENTS OF DECISION
MAKING
1. Decision Alternatives :-
These are the finite number of options

available with the decision maker to solve
the business problem.
Number and type of alternatives depend

on the previous decisions made and their
outcomes.
MAKING
These decision alternatives are in the
control of the decision maker.
The decision alternatives may be

numerically or non-numerically
described.
MAKING
2. States of nature
It is an event or scenario that is not under

the control of the decision maker.
It may be a state of economy, a weather

condition, a political development, etc.
MAKING
It can be identified by scenario analysis
through interviews of – stakeholders, long
time managers etc. to understand states of
nature that may have serious impact on
the decision.
States of nature are mutually exclusive

and collectively exhaustive w.r.t. any
decision problem.
MAKING
3. Payoff
Itis the numerical value (outcome which is

generally in monetary units) obtained due to
the application of each possible combination
of decision alternatives and states of nature.
The payoffs are measured within a specified

period (e.g. year, month, etc) called as
decision horizon.
STEPS OF DECISION
MAKING PROCESS
1. Identify and define the problem.
2. List all possible

future events (not
under the control
of decision-maker)
that are likely to
occur.
STEPS OF DECISION
MAKING PROCESS
3. Identify all the courses of action available
to the decision maker.
4. Express the payoffs (Pij) resulting from

each combination of course and state of
nature.
5. Apply an appropriate model to select

the best course of action on the basis of
criterion to get optimal payoff.
DEGREE OF KNOWLEDGE
Decision theory is a descriptive and prescriptive
business modeling approach to classify the degree
of knowledge and compare expected outcomes
due to several courses of action.
The degree of Knowledge is divided into 4

categories :-
Ignorance
Uncertainty
Risk
Certainty
TYPES OF DECISION MAKING
ENVIRONMENTS
1. Decision Making under Certainty
- Perfect information and knowledge of the outcome of
each decision alternative.
2. Decision Making under Risk

- Probability of occurrence
of events is known.
3. Decision Making under

Uncertainty
- Probability of occurrence
of events is unknown.
EXAMPLE I
 A firm Manufactures 3 types of products. The fixed and
variable costs are given below :-
Product Type Fixed Cost (Rs) Variable Cost
(Rs / unit)
Product A 25000 12
Product B 35000 9
Product C 53000 7
The Likely demand (units) of the products is given below:-

Demand No. of Units
Poor 3000
Moderate 7000
High 11000
If the sale price of each type of product is Rs 25, then

prepare the payoff matrix.
EXAMPLE I SOLUTION
LetD1, D2 and D3 be the event of poor,
moderate and high demand.
The payoff in this case is determined by

Payoff = Sales revenue – Cost (fixed +
Variable)
So, the payoff for each pair of demand (state

of nature or event) and type of product
(action or alternative) is given by :-
EXAMPLE I SOLUTION
D1A = (3000 x 25) – 25000 – (3000 x 12)
= 14000
D1B = (3000 x 25) – 35000 – (3000 x 9)

= 13000
.
.
D3C = (11000 x 25) – 53000 – (11000 x 7)
= 145000
EXAMPLE I SOLUTION
• The Payoff Matrix is given by
Product D1 D2 D3
Type
A 14000 66000 118000
B 13000 77000 141000
C 1000 73000 145000

EXAMPLE I SOLUTION
• If payoffs are expressed in (‘000 Rs) then
the payoff matrix is given by
Product D1 D2 D3
Type
A 14 66 118
B 13 77 141
C 1 73 145
DECISION MAKING UNDER
UNCERTAINITY
The decision situations where there is no way in
which the decision maker can assess the
probabilities of the
various states of nature are
called decisions under
uncertainty.
When the probability of

the outcome cannot be
quantified, the decision maker must arrive at a
decision based on actual conditional payoff
keeping in view the criterion of effectiveness.
UNCERTAINITY
With the probabilities of various outcomes
unknown, the actual decisions are based on
specific criteria which are as follows :-
1. Optimism (Maximax or minimin) Criterion
2. Pessimism (Maximin or minimax) Criterion
3. Equal Probabilities (Laplace) Criterion
4. Coefficient of Optimism (Hurwicz) Criterion
5. Regret (Savage) or Opportunity loss Criterion

UNCERTAINITY
Sr. Decision making Criterion to be Procedure
No. under uncertainity used
a) Find the maximum (or

Minimum in case of cost)
payoff in each of the action.
Maximax (Or
1 Optimism Criterion Minimin in case b) Select the maximum (or
of cost) Minimum in case of cost) value
from these.
c) The action corresponding to this

value is the Optimal action
UNCERTAINITY
a) Find the minimum (or

Maximum in case of cost)
payoff in each of the action.
2 Pessimism (Maximin Maximin (or
or Minimax) Criterion Minimax in case b) Select the maximum (Minimum
of cost) in case of cost) value from
these.
c) The action corresponding to

this value is the Optimal action
UNCERTAINITY
a) Calculate the average of all the

payoffs of each action.
Maximum (or
3 Equal Probabilities Minimum in b)Select the maximum (or
(Laplace) Criterion case of cost) minimum in case of cost) value
Mean or out of these.
Average
c)The action corresponding to this
UNCERTAINITY
a) Find the maximum payoff in
each of the action.
b) Find the minimum payoff in
α (max) + (1 – each of the action.
Coefficient of α) (min) c) Calculate the Hurwicz criterion
4 Optimism (Hurwicz) value of all the payoffs of each
Criterion [OR α (min) + action.
(1 – α) (max) in d) Select the maximum (or
case of cost] minimum in case of cost) value
from these.
e) The action corresponding to this
Where, α = Coefficient of Optimism

• Its value lies between 0 and 1
• 0 denotes pessimistic attitude whereas 1 denotes optimistic attitude
UNCERTAINITY
a) Prepare a new regret table by

subtracting all other values
payoff in each event from
highest payoff of the
corresponding event
5 Regret (Savage) Minimax regret
Criterion b) Find the maximum regret in
each of the action.
c) Select the minimum value from

these.
d) The action corresponding to this

EXAMPLE II
The following profit matrix gives the
payoff (in Rs) for 3 different strategies
(Expand, Construct and subcontract)
against 4 events of demand (high,
moderate, low and nil) for a particular
company.
Strategies Events (Demand)
High Moderate Low Nil
Expand 50000 25000 -25000 -45000
Construct 70000 30000 -40000 -80000
Subcontract 30000 15000 -1000 -10000

EXAMPLE II
Indicate the optimal decision taken under the
following approaches:-

(α = 0.8)

EXAMPLE III
 A manufacturer makes a product, of which the principal
ingredient is a chemical X. At present the manufacturer spends
Rs 1000 per year on supply of X, but there is a possibility that
its price may rise to four times its present value due to
shortage. There is another chemical Y which the manufacturer
could use in conjunction with a third chemical Z in order to
give the same effect as chemical X. Chemicals Y and Z would
together cost the manufacturer Rs 3000 per year and their
prices are not likely to increase. What action should the
manufacturer take based on following approaches :-
1. Pessimism Criterion
2. Regret Criterion
3. Hurwicz Criterion (α = 0.4)
EXAMPLE III SOLUTION
Profit Matrix
Events Acts
Use X Use Y & Z
Price of X -4000 -3000
increases
Price of X does not -1000 -3000
increase
EXAMPLE III SOLUTION
Cost (loss) Matrix

Events Acts
Use X Use Y & Z
Price of X 4000 3000
increases
Price of X does not 1000 3000
increase
EXAMPLE IV
A bookstore sells a particular book of tax
laws for Rs100. He purchases the book
for Rs 80 per copy. Since some tax laws
change every year, the copies unsold at
the end of a year become outdated and
can be disposed off for Rs 30 each.
According to past experience, the annual
demand for this book is between 18 and
23 copies. Assuming that the order for this
book can be placed only once during the
year, the
EXAMPLE IV
problem before the store manager is to decide
optimal number of books that should be purchased
for the next year based on following criterion :-

EXAMPLE IV SOLUTION
There are 6 possible events :-
E1 = 18 copies are demanded
EXAMPLE IV SOLUTION
There are 6 possible strategies or courses
of action :-
A1 = Buy 18 copies
A2 = Buy 19 copies
A3 = Buy 20 copies
A4 = Buy 21 copies
A5 = Buy 22 copies
A6 = Buy 23 copies
EXAMPLE IV SOLUTION
The Profit function will be :-
Let D be the demand in units
And S be the quantity decided to purchase
Then,
P = 20S …when D > S
P = 20D – 50 (S – D) …when D < S

= 70D – 50S
EXAMPLE IV SOLUTION
Pay off Table :-
Event, Ei Act, Aj (Supply)
(Demand)
A1 : 18 A2 :19 A3 : 20 A4 : 21 A5 : 22 A6 : 23
E1 : 18 360 310 260 210 160 110

E2 : 19 360 380 330 280 230 180
E3 : 20 360 380 400 350 300 250
E4 : 21 360 380 400 420 370 320
E5 : 22 360 380 400 420 440 390
E6 : 23 360 380 400 420 440 460
EXAMPLE IV SOLUTION
1. Optimism (Maximax or Minimin) Criterion :-
(Demand)
A1 : 18 A2 :19 A3 : 20 A4 : 21 A5 : 22 A6 : 23
E1 : 18 360 310 260 210 160 110

E2 : 19 360 380 330 280 230 180
E3 : 20 360 380 400 350 300 250
E4 : 21 360 380 400 420 370 320
E5 : 22 360 380 400 420 440 390
E6 : 23 360 380 400 420 440 460
Maximum 360 380 400 420 440 460
Payoff
Maximum
EXAMPLE IV SOLUTION
2. Pessimism (Maximin or Minimax)
Criterion :-
(Demand)
A1 : 18 A2 :19 A3 : 20 A4 : 21 A5 : 22 A6 : 23
E1 : 18 360 310 260 210 160 110

E2 : 19 360 380 330 280 230 180
E3 : 20 360 380 400 350 300 250
E4 : 21 360 380 400 420 370 320
E5 : 22 360 380 400 420 440 390
E6 : 23 360 380 400 420 440 460
Minimum 360 310 260 210 160 110
Payoff
Maximum
EXAMPLE IV SOLUTION
3. Equal Probability (Laplace) Criterion :-
(Demand)
A1 : 18 A2 :19 A3 : 20 A4 : 21 A5 : 22 A6 : 23
E1 : 18 360 310 260 210 160 110

E2 : 19 360 380 330 280 230 180
E3 : 20 360 380 400 350 300 250
E4 : 21 360 380 400 420 370 320
E5 : 22 360 380 400 420 440 390
E6 : 23 360 380 400 420 440 460
Mean 360 368.3 365 350 323.3 285
(Expected)
Payoff
Maximum
EXAMPLE IV SOLUTION
4. Hurwicz Criterion :- (assume α = 0.6)
(Demand) A1 : 18 A2 :19 A3 : 20 A4 : 21 A5 : 22 A6 : 23
E1 : 18 360 310 260 210 160 110

E2 : 19 360 380 330 280 230 180
E3 : 20 360 380 400 350 300 250
E4 : 21 360 380 400 420 370 320
E5 : 22 360 380 400 420 440 390
E6 : 23 360 380 400 420 440 460
Maximum 360 380 400 420 440 460
Payoff
Minimum 360 310 260 210 160 110
Payoff
α (max) + (1 360 352 344 336 328 320
– α) (min)
Maximum
EXAMPLE IV SOLUTION
5. Regret (Savage) or Opportunity Loss criterion :-
REGRET TABLE

(Demand)
A1 : 18 A2 :19 A3 : 20 A4 : 21 A5 : 22 A6 : 23
E1 : 18 0 50 100 150 200 250

E2 : 19 20 0 50 100 150 200
E3 : 20 40 20 0 50 100 150
E4 : 21 60 40 20 0 50 100
E5 : 22 80 60 40 20 0 50
E6 : 23 100 80 60 40 20 0
Maximum 100 80 100 150 200 250
Regret Minimum
EXAMPLE V
XYZ informatics summarizes international
financial information reports on weekly basis,
prints sophisticated data and forecasts, which
are purchased weekly by mutual funds, banks
and insurance companies. This information is
very expensive and demand for the report is
limited to a maximum of 30 units per week. The
possible demands are 0,10,20, or 30 per week.
The profit per report sold is Rs 30 and the loss
per report unsold is Rs 20. No production of
extra reports during a week is possible.
EXAMPLE V
Further there is a penalty cost of Rs 250 for not meeting
the demand. Unsold reports cannot be carried over to next
week. Find the number of reports to be produced based on
following criterion :-
4. Coefficient of Optimism (Hurwicz) Criterion (ɑ = 0.8)

DECISION MAKING UNDER RISK
 The decision situations wherein the decision maker
chooses to consider several possible outcomes and the
probabilities of their occurrence can be stated are called
decision under risk.
 The decision maker has

sufficient information for
assigning probability for
each state of nature
(event)
 After assigning probabilities the decision maker needs to

select the action (alternative) which will have maximum
expected payoff.
 The probabilities of various outcomes may be
determined objectively from the past records.
 Pastrecords may not

be available in many
cases to arrive at
objective probabilities.
 In such cases the decision maker may, on the basis

of his experience and judgment, be able to assign
subjective probabilities, and the problem can be
solved as decision making under risk.
1. Maximum Likelihood Principle :-
Under this principle, the decision maker first

considers the event that is most likely to occur i.e.
which has the highest probability. He then selects
the action which has maximum (or minimum in
case of cost) payoff corresponding to the event
which has maximum probability.
EXAMPLE V
 A retailer purchases cherries every morning at Rs
50 per case and sells them for Rs 80 per case. Any
case that remains unsold at the end of the day can
be disposed off the next day at salvage value of Rs
20 per case. Past sales have ranged from 15 to 18
cases per day. The following is the record of sales
for past 120 days.
Cases Sold 15 16 17 18
Number of days 12 24 48 36
Find out how many cases per day should the

retailer purchase in order to maximize
his profit using maximum likelihood principle.
EXAMPLE V SOLUTION
Conditional Payoff (Profit) Matrix :-
States of Probability Conditional Payoff (Profit) due to course of action
Nature or (Supply per day)
Event Buy 15 Buy 16 Buy 17 Buy18
(Demand
per day)
15 0.1 450 420 390 360

16 0.2 450 480 450 420
17 0.4 450 480 510 480
18 the event
Since, 0.3 of having
450 demand
480 of 17 cases
510 is having
540
maximum probability i.e. 0.4, therefore, the optimal course
of action will be to buy 17 cases per day to get maximum
expected payoff of 510 Rs
2. Expectation Principle :-
 Generally the decision making in situations of risk is on

the basis of expectation principle.
a) Expected Payoff or Expected Monetary Value

(EMV) :- With the event probabilities assigned
objectively or subjectively, the expected payoff for
each strategy is calculated by multiplying the payoff
values with their respective probabilities and then
adding up the products. The strategy with the highest
(or lowest) payoff represents the optimal choice.
 The expected monetary value (EMV) for a given course
of action is obtained by summation of the payoff values
multiplied by the probabilities associated with each state
of nature.
m
 EMV (for any action Sj) = ∑ Pij . pi
i=1
Where ,
m = number of states of nature (events)
pi = probability of occurrence of state of nature, Ni
Pij = payoff associated with state of nature Ni and
course of action Sj
PROCEDURE FOR CALCULATING EMV
1. Construct a matrix listing all possible courses of
action (strategies) and states of nature (events).
2. Enter the conditional payoffs values for each course of

action and state of nature combination. Enter the
probabilities of each state of nature (event).
3. Calculate the EMV for each course of action by

multiplying the conditional payoffs by associated
probabilities and adding these values for each course
of action.
4. Select the optimal course of action which gives

maximum (or minimum in case of cost) EMV
EXAMPLE V
 A retailer purchases cherries every morning at Rs 50 per
case and sells them for Rs 80 per case. Any case that
remains unsold at the end of the day can be disposed off
the next day at salvage value of Rs 20 per case. Past sales
have ranged from 15 to 18 cases per day. The following
is the record of sales for past 120 days.

retailer purchase per day in order to maximize
his profit using Expectation principle or EMV method.
EXAMPLE V SOLUTION
(Demand
per day)
15 0.1 450 420 390 360

16 0.2 450 480 450 420
17 0.4 450 480 510 480
18 0.3 450 480 510 540
EXAMPLE V SOLUTION
Expected Payoff (Profit) Matrix :-
States of Probability Expected Payoff (Profit) due to course of action
(Demand
per day)
15 0.1 45 42 39 36
16 0.2 90 96 90 84
17 0.4 180 192 204 192
18 0.3 135 144 153 162
Expected Monetary Value 450 474 486 474
(EMV) Maximum EMV
b) Expected Opportunity Loss (EOL) or Expected Regret :-
 The Expected Opportunity Loss (EOL) or expected regret is

defined as the difference between the highest Payoff (profit) and
the actual payoff (profit) due to choosing a particular course of
action in a particular state of nature.
 EOL or expected regret is the amount of payoff that is lost by not

choosing a course of action resulting to the minimum payoff in a
particular state of nature.
 The EOL or expected regret for any course of action is determined

by summing up the products of the regret value and their
corresponding probabilities. The optimal strategy is the one
which minimises expected regret.
m
 EOL (for any action Sj) = ∑ Lij . pi
i=1
Where,
Lij = opportunity loss due to state of nature Ni and
course of action Sj
1. Prepare a conditional payoff matrix for each
combination of course of action and state of nature.
2. For each state of nature calculate the conditional

opportunity loss or regret by subtracting each payoff
from maximum payoff.
3. Calculate the EOL or expected regret by multiplying

the probability of each state of nature with conditional
regret and then adding these values.
4. Select a course of action for which EOL or expected

regret is minimum.
EXAMPLE V

his profit using Expected Opportunity Loss (EOL)
principle or Expected regret method.
EXAMPLE V SOLUTION
(Demand
per day)
15 0.1 450 420 390 360

16 0.2 450 480 450 420
17 0.4 450 480 510 480
18 0.3 450 480 510 540
EXAMPLE V SOLUTION
Conditional Opportunity Loss (Regret) Matrix :-
States of Probability Conditional Opportunity Loss or regret due to
Nature or course of action (Supply per day)
(Demand
per day)
15 0.1 0 30 60 90
16 0.2 30 0 30 60
17 0.4 60 30 0 30
18 0.3 90 60 30 0
EXAMPLE V SOLUTION
Expected Opportunity Loss (EOL) or Expected regret Matrix :-
States of Probability Expected Opportunity Loss(EOL) due to course of
Nature or action (Supply per day)
(Demand
per day)
15 0.1 0 3 6 9
16 0.2 6 0 6 12
17 0.4 24 12 0 12
18 0.3 27 18 9 0
Expected Opportunity Loss 57 33 21 33
(EOL) or Expected Regret Minimum
EXPECTED VALUE OF PERFECT
INFORMATION (EVPI)
If the decision makers can get perfect
information about the occurrence of various
states of nature, then choosing a course of action
that gives the desired payoff in presence of any
state of nature becomes easy.
Expected value of perfect information (EVPI)

represents maximum amount of money required
to pay for getting additional information about
the occurrence of various states of nature before
arriving a decision.
INFORMATION (EVPI)
Expected value of perfect information =
Expected Payoff (profit) with perfect
information – Expected Payoff (profit)
without perfect information
 EVPI = EPPI – max EMV

INFORMATION (EVPI)
m
 EVPI = ∑ pi.(max Pij) – max EMV
i=1
Where,
max Pij = best payoff when action, Sj is taken in the
presence of state of nature Ni
max EMV = Maximum Expected monetary value
EXAMPLE V

his profit using Expectation principle or EMV method.
What will be the maximum amount that the retailer
could be willing to pay to get perfect information (EVPI).
EXAMPLE V SOLUTION
(Demand
per day)
15 0.1 450 420 390 360

16 0.2 450 480 450 420
17 0.4 450 480 510 480
18 0.3 450 480 510 540
EXAMPLE V SOLUTION
(Demand
per day)
15 0.1 45 42 39 36
16 0.2 90 96 90 84
17 0.4 180 192 204 192
18 0.3 135 144 153 162
Expected Monetary Value 450 474 486 474
(EMV) Maximum EMV
EXAMPLE V SOLUTION
(Demand
per day)
15 0.1 450
16 0.2 480
17 0.4 510
18 0.3 540
Expected Payoff of Perfect 507
Information
EVPI = EPPI (EPPI)
– Max EMV
= 507 - 486
= 21
EXAMPLE VI
A toy manufacturer is considering a project of
manufacturing a dancing doll with 3 different
movement designs. The doll will be sold at an
average price of Rs 10. The first movement
design using ‘gears and levels’ will have a fixed
cost of Rs 100000 and variable cost of Rs 5 per
unit. A second design with spring action will
have a fixed cost of 160000 and variable cost of
Rs 4 per unit. Another third design with
‘weights and pulleys’ will have a fixed cost of
Rs 300000 and variable cost of Rs 3 per unit.
EXAMPLE VI
The demand events that can occur for the doll and
the probabilities is given below :-
Demand No. of Units Probability
Light 25000 0.10
Moderate 100000 0.70
High 150000 0.20
a) Construct a payoff table

b) Find the optimal decision using EMV method
c) Find the optimal decision using EOL method
d) How much can the toy manufacturer afford to
pay in order to obtain perfect information
about the demand ?
EXAMPLE VII
A TV dealer finds that the cost of a TV in stock
for a week is Rs 30 and the cost of a unit
shortage is Rs 70. For one particular model of
TV the probability distribution is as follows:-
Weekly Sales 0 1 2 3 4 5 6
Probability 0.10 0.10 0.20 0.25 0.15 0.15 0.05
Find out how many units per week should the

dealer order using i) EMV Method ii) EOL
method iii) Also find EVPI
EXAMPLE VII
A Company manufactures parts for passenger
cars and sells them in lots of 10000 parts. The
company has the policy of inspecting each lot
before it is shipped to the customer. Five
inspection ratings established for quality control
represent the percentage of defective items
contained in each lot. The management is
considering two possible courses of action :-
i) Shut down the entire plant operations and
thoroughly inspect each machine. This action
costs Rs 600
EXAMPLE VII
ii) Continue production but offer the customer
a refund of Re 1 for each defective item that is
detected and returned.
Rating Proportion of defectives Frequency
Excellent (A) 0.02 25
Good (B) 0.05 30
Acceptable (C) 0.10 20
Fair (D) 0.15 20
Poor (E) 0.20 5
Find out the optimal decision for the company under

i) EMV Method ii) EOL method iii) Also find EVPI
EXAMPLE VIII
A company has recently installed new
machinery but has not yet decided on the
appropriate number of a certain spare part
required for repairs.
Spare part costs Rs 2000 each but are only
available if ordered now. If the plant failed and
there was no spare part available, the cost to the
business mending the plant is Rs 15000 per
failure. The plant has an estimated life of 10 years
and the probability distribution of failures during
this time, based on experience with similar plants,
is as follows :-
EXAMPLE VIII
No. of 0 1 2 3 4 5 and
Failures / above
10 years
Probability 0.1 0.4 0.3 0.1 0.1 nil
Calculate :-
a) Expected number of failures in a 10 year period
b) The Optimal number of spares to be purchased
using i) EMV method ii) EOL method
c) The cost of ordering in i) EMV method ii) EOL
method
d) The Expected value of perfect information of
number of failures in a 10 year life
EXAMPLE IX
A company has total excess cash funds of Rs
60000 to invest in various projects during this
month and next, according to the cash flow
statement prepared by the accounts department.
The company has been offered the following
investment opportunities : It can participate
immediately (at the start of the month) in a
project by investing 60000, which is equally
likely to result in a net profit of Rs 20000 or a
loss of Rs10000 within the month. In effect, the
company will be able to reclaim its principal,
with either profit or loss, by the
EXAMPLE IX
months end. At the same time the company is
informed that in one month from now, it will be
given an opportunity of investing Rs 55000 in
another investment which is equally likely to result
in a net profit of Rs 15000 or a net loss of Rs 5000.
Assuming that this company examines its cash
position every 2 months to determine the feasibility
of investing excess cash, advise this company as to
whether it should invest in project 1 of this month or
invest in project 2 of next month or invest in
projects 1and 2 together if the objective is to
maximise profits over the next 2 months.
EXAMPLE X
A bookstore sells a particular book of tax
laws for Rs100. He purchases the book
for Rs 80 per copy. Since some tax laws
change every year, the copies unsold at
the end of a year become outdated and
can be disposed off for Rs 30 each.
According to past experience, the annual
demand for this book is between 18 and
23 copies. Assuming that the order for this
book can be placed only once during the
year, the
EXAMPLE X
problem before the store manager is to decide
optimal number of books that should be purchased
for the next year based on following criterion :-
i) EMV Method ii) EOL method iii) Also find

EVPI
The probabilities of the events are given as
follows:- E1 = 0.05, E2 = 0.10, E3 = 0.30, E4 =
0.40, E5 = 0.10, E6 = 0.05
EXAMPLE I SOLUTION
a) Using Maximum Likelihood Principle
Event, Probab Act, Aj
Ei ility Pi
A1:18 A2:19 A3:20 A4:21 A5:22 A6:23
E1: 18 0.05 360 310 260 210 160 110

E2: 19 0.10 360 380 330 280 230 180
E3: 20 0.30 360 380 400 350 300 250
E4: 21 0.40 360 380 400 420 370 320
E5: 22 0.10 360 380 400 420 440 390
E6: 23 0.05 360 380 400 420 440 460
Here, the event most likely to occur is Event E4 (p = 0.40)

and the maximum conditional payoff corresponding to it is
420. Hence the optimal strategy under maximum
likelihood principle is A4
EXAMPLE I SOLUTION
b) Using Expectation principle
Ei ility Pi
A1:18 A2:19 A3:20 A4:21 A5:22 A6:23
E1: 18 0.05 360 310 260 210 160 110

E2: 19 0.10 360 380 330 280 230 180
E3: 20 0.30 360 380 400 350 300 250
E4: 21 0.40 360 380 400 420 370 320
E5: 22 0.10 360 380 400 420 440 390
E6: 23 0.05 360 380 400 420 440 460
Expected Payoff 360 376.5 386 374.5 335 288.5
Here, the maximum expected payoff or expected monetary
value is 386, Hence the optimal strategy is A3.
EXAMPLE I SOLUTION
Using Expected Opportunity Loss or Expected Regret
Ei ility Pi
A1:18 A2:19 A3:20 A4:21 A5:22 A6:23
E1: 18 0.05 0 50 100 150 200 250

E2: 19 0.10 20 0 50 100 150 200
E3: 20 0.30 40 20 0 50 100 150
E4: 21 0.40 60 40 20 0 50 100
E5: 22 0.10 80 60 40 20 0 50
E6: 23 0.05 100 80 60 40 20 0
Expected Regret 51 34.5 25 36.5 76 122.5
Here, the minimum expected regret or expected
opportunity loss is 25, Hence the optimal strategy is A3.
EXAMPLE I SOLUTION
Expected Payoff of Perfect Information (EPPI)
Ei ility Pi
A1:18 A2:19 A3:20 A4:21 A5:22 A6:23
E1: 18 0.05 360

E2: 19 0.10 380
E3: 20 0.30 400
E4: 21 0.40 420
E5: 22 0.10 440
E6: 23 0.05 460
EPPI 411
EXAMPLE I SOLUTION
Expected Value of Perfect Information
(EVPI)
EVPI = EPPI – Max EMV
EVPI = 411- 386
EVPI = 25
DECISION TREE
In single stage decision making the
payoffs, actions, states of nature and
associated probabilities were not
subjected to changes.
There are situations that involve multiple

stages which are characterised by a
sequence of decisions with each decision
influencing the next.
DECISION TREE
Such a situation is called sequential or
multiperiod decision process.
 E.g. Test marketing of a new product
A decision tree analysis involves the

construction of a diagram that shows, at a
glance, when decisions are expected to be
made – in what sequence, their possible
outcomes and the corresponding payoffs.
DECISION TREE
A decision tree consists of the
following :-
1. Nodes
2. Branches
3. Probability estimates
4. Payoffs
TYPES OF NODES
1. Decision (or action) Node :-
It is represented by a square and indicates a point of
time where the decision maker must select one course
of action amongst the available. The courses of action
are the branches emerging out of decision node.
2. Chance (or event) node :-

It is represented by a circle and indicates a point of
time where the decision maker will discover the
response to his decision. The states of nature are the
branches emerging out of chance node.
TYPES OF BRANCHES
1. Decision Branch :-
It is the branch coming out from a decision node and
represents a course of action that can be chosen at a
decision point.
2. Chance Branch :-
It is the branch coming out from a chance node and
represents a state of nature resulting from a course of
action.
3. Terminal Branch :-
A branch that forms the end of the decision tree i.e. not
followed by a decision or chance node is called as terminal
branch.
DECISION TREE
The general approach used in decision
tree analysis is to work backward through
the tree from right to left, computing the
expected value or EMV (also called
position value) of each chance node.
We then choose a particular branch

(action) of decision node which leads to
chance node with highest expected payoff
or EMV.
STEPS IN DECISION TREE
ANALYSIS
1. Identify the decision points (or decision nodes) and
courses of action at each decision point.
2. Identify the chance nodes and the states of nature at

each chance nodes
3. Commencing from the extreme right, Determine the

expected payoff (or EMV) at each chance node by
summing the products of the probabilities and payoff
from the branches (events).
4. Determine the expected payoff (or EMV) at each

decision node by selecting the highest payoff from its
branches (actions).
STEPS IN DECISION TREE
ANALYSIS
5. Proceed backwards to next stage of decision points
(or decision nodes).
6. Repeat the above steps till the first decision point is

reached.
7. Finally, identify the courses of action to be adopted

from beginning to the end under different possible
outcomes for the situation as a whole.
EXAMPLE XI
Show the decision situation in the form of
a decision tree and indicate the most
preferred decision and its corresponding
expected value for the following table :-
Demand Probability Supply (Actions)
(Events)
Produce 25 Produce 75
units units
High 0.6 4000 10000
Low 0.4 2000 -5000
EXAMPLE XI SOLUTION
Action 1 & Event 1
Event 1 Payoff
A
Action 1 Action 1 & Event 2
Event 2
Payoff
1
Action 2 & Event 1

Action 2 Event 1 Payoff
Event 2 Action 2 & Event 2

Payoff
EXAMPLE XI SOLUTION
High (0.6) Rs 4000
EMV A =
(Event)
Rs 3200
A
EMV 1 = Produce 25
Rs 4000 (Action) Low (0.4)
(Event) Rs 2000
1
High (0.6) Rs 10000

Produce 75 (Event)
(Action)
B
EMV B =
Rs 4000 Low (0.4)
(Event) Rs -5000
EXAMPLE XI SOLUTION
4000
0.6
3200
Produce 25
0.4
2000
4000
10000
Produce 75 0.6
4000
0.4
-5000
EXAMPLE XI SOLUTION
 Evaluation of Decision Node 1
Events
Conditional Expected
Actions Probability
Payoff Payoff
(Demand)
1. Produce 25 High 0.6 4000 2400
Low 0.4 2000 800
Total EMV 3200
2. Produce 75 High 0.6 10000 6000
Low 0.4 -5000 -2000
Total EMV 4000
The optimal decision is “to produce 75 units” so as to get a

maximum EMV of 4000 Rs.
EXAMPLE XII
 An oil company has recently acquired rights in a certain
area to conduct surveys and test drillings to lead to lifting
oil if it is found in commercially exploitable quantities.
The area is considered to have good potential for
finding oil in commercial quantities. At the outset, the
company has the choice to conduct further geological
tests or to carry out a drilling programme immediately.
On the known conditions, the company estimates that
there is a 70 : 30 chance of further tests showing a
‘success’.
Whether the tests show the possibility of ultimate
success or not or even if no tests are undertaken at all, the
company could still pursue its drilling programme or
alternatively consider selling its rights to drill in the area.
EXAMPLE XII
Thereafter, however, if it carries out the drilling
programme, the likelihood of final success or failure is
considered dependent on foregoing stages which are as
follows :-
 If ‘successful’ tests have been carried out, the
expectation of success in drilling is given as 80 : 20.
 If the tests indicate ‘failure’, then the expectation of
success in drilling is given as 20 : 80.
 If no tests have been carried out at all, the expectation of
success in drilling is given as 55 : 45.
Costs and revenues have been estimated for all possible
outcomes and the net present value of each is as follows :-
EXAMPLE XII
Net Present Value
Outcome
(Rs Million)
With prior tests 100
Success
Without prior tests 120
With prior tests -50
Failure
Without prior tests -40
Prior tests show ‘success’ 65
Sale of exploitation
rights Prior tests show ‘failure’ 15
Without prior tests 45
i) Represent the above information using a decision tree.

ii) Evaluate the tree in order to advise the management
of the company on its best course of action.
EXAMPLE XII SOLUTION
Failure (0.45)
-40
(Event) 65 -50
Sell
Failure (0.20)
(Action)
C Success (0.55)
120 (Event)
Drill (Event)
(Action)
2 B Success (0.80)
100
Positive (0.70) Drill
(Event) (Action) (Event)
3 D
Test
(Action)
Negative (0.30) Drill Success (0.20)
(Event) (Action) (Event)
Sell
1 A 100
(Action)
45 Failure (0.80)
Sell (Event)
(Action)
15 -50
-40
Failure (0.45) 65 -50
Sell Failure (0.20)
48
Success (0.55)
120
Drill
70 70
Success (0.80)
100
Positive (0.70) Drill
53.5 53.5
Test
Negative (0.30)
Drill
Success (0.20)
Sell
15 -20 100
45 Failure (0.80)
Sell
15 -50
EXAMPLE XIII
A Finance Manager is considering drilling a well.
In the past, only 70% of wells drilled were
successful at 20 metres depth in that area.
Moreover, on finding no water at 20 m, some
persons in that area drilled it further up to 25 m but
only 20% struck water at that level. The prevailing
cost of drilling is Rs 500/m.
The Finance Manager estimated that in case
he does not get water in his own well, he will
have to pay Rs 15000 to buy water from outside
for the same period of getting water from the well.
EXAMPLE XIII
The following decisions are considered :-
i) Do not drill any well;
ii) Drill up to 20 m, and
iii) If no water is found at 20 m, drill further up to
25 m.
Draw an appropriate decision tree and determine
the Finance Manager’s optimal strategy.
EXAMPLE XIII SOLUTION
27500
No water (0.80)
15000
Do not Drill A
Drill up to 25 m
water (0.20)
2 1
12500
No water (0.30)
Drill up to 20 m Do not Drill
B
25000
water (0.70)
10000
EXAMPLE XIII SOLUTION
27500
No water (0.80)
15000
Do not Drill 24500
Drill up to 25 m
water (0.20)
14350
24500
12500
No water (0.30)
Drill up to 20 m Do not Drill
14350
25000
water (0.70)
10000
EXAMPLE XIV
 An investor has two independent investments A and B
available to him but he lacks the capital to undertake
both of them simultaneously. He can choose to take A
first and then stop or if A is successful then take B or
vice versa. The probability of success on A is 0.8 while
for B is 0.4. Both investments require an initial capital
outlay of Rs 20000 and both return nothing if the venture
is unsuccessful. Successful completion of A will return
Rs 30000 over cost while successful completion of B
will return Rs 60000 over cost.
i) Draw a decision tree and determine the best strategy
ii) Solve the problem by preparing a payoff table
EXAMPLE XIV SOLUTION
Failure (0.20) -20000 Success (0.40) 60000
C A
Accept B
Success (0.8) Failure (0.60) -20000

Accept A 1
Stop
0
Invest in none
3 0
Accept B
Failure (0.60) -20000 Success (0.80) 30000
D B
Accept A
Success (0.40) Failure (0.20) -20000

2
Stop
0
Failure (0.20) -20000 Success (0.40) 60000
29600 12000
Accept B
(30000 + 12000)
Success (0.8) Failure (0.60) -20000

Accept A 42000
Stop 0
Invest in none
29600 0
Accept B
Failure (0.60) -20000 Success (0.80) 30000
20000 20000
Accept A
Success (0.40) Failure (0.20) -20000

80000
Stop
(60000 + 20000) 0
 Courses of action:
A1: Do not invest
A2: Accept A and then stop
A3: Accept B and then stop
A4: Accept A and, if successful, Accept B
A5: Accept B and, if successful, Accept A
 Events:
E1: Both A and B are successful (0.8 X 0.4)
E2: A is successful but not B (0.8 X0.6)
E3: B is successful but not A (0.2 X 0.4)
E4: Neither A nor B is successful (0.2 X 0.6)
 Conditional Payoff Matrix:-
Course of Action
Event Probability
A1 A2 A3 A4 A5
E1 0.32 0 30000 60000 90000 90000
E2 0.48 0 30000 -20000 10000 -20000
E3 0.08 0 -20000 60000 -20000 40000
E4 0.12 0 -20000 -20000 -20000 -20000

 Expected Payoff Matrix:-
Course of Action
Event Probability
A1 A2 A3 A4 A5
E1 0.32 0 9600 19200 28800 28800
E2 0.48 0 14400 -9600 4800 -9600
E3 0.08 0 -1600 4800 -1600 3200
E4 0.12 0 -2400 -2400 -2400 -2400
Expected Pay off or

0 20000 12000 29600 20000
EMV
EXAMPLE XV
 An oil company has to decide whether to drill for oil at a
newly discovered oil field. It would cost the company Rs
400 crore to go ahead with the project and if oil is found
the value is estimated to be Rs 1600 crore. The company
believes that there is 40% chance that the field contains oil.
 Before drilling, the company has the option to hire a
consultant geologist to obtain more information about the
likelihood that the field will contain oil. The consultant
may give one of the following two reports – favourable or
unfavourable. It is believed that the chances of favourable
and unfavourable report are 60 : 40.
EXAMPLE XV
 A favourable report would make the company
believe for an 80% chance of the field containing oil
while an unfavourable report would cause the
probability of discovering oil to be revised
downwards to 10%.
 If hiring the geologist would cost Rs 25 crore,
determine optimal course of action for the company.
Draw the decision tree diagram to solve the problem
and also determine EVSI and EVPI.
Oil (0.40)
1200
(Event) 0 -400
Do not Drill
(Action) No Oil (0.20)
C No Oil (0.60)
-400 (Event)
Drill (Event)
(Action) Favourable 1 A Oil (0.80)
1200
Drill
(0.60) (Event) (Event)
(Action)
3 Engage
D
Geologist
(Action)
Unfavourable Oil (0.10)
Drill
(Action)
Do not Drill 2 B 1200
0 No Oil (0.90)
(Event)
Do not Drill
(Action)
0 -400
Oil (0.40)
1200
(Event) 0 -400
Do not Drill
(Action) No Oil (0.20)
240
No Oil (0.60)
-400 (Event)
Drill (Event)
(Action) Favourable 855
880
Oil (0.80)
1200
Drill
(880 - 25) (Action)
503 Engage
503
Geologist
(Action) Drill
Unfavourable (0 - 25) Oil (0.10)
(Action)
-25
-240
1200
Do not Drill
0 No Oil (0.90)
(Event)
Do not Drill
(Action)
0 -400
EXAMPLE XVI
 A company which has recently invented a telephone device
is faced with the problem of selecting out of the following
courses of action available:
i. Manufacture the device itself
ii. Allow production on royalty basis by another
manufacturer; or
iii. Sell the rights for its invention for a lump sum.
The profit (in ‘000 Rs) expected in each case and the
probabilities associated with the level of sales are shown in
the following table:
Action
Outcome Probability
Manufacture Royalties Sell all rights
High Sales 0.2 800 350 180
Medium Sales 0.3 300 200 180
Low Sales 0.5 -100 100 180
EXAMPLE XVI
 Represent the company’s problem in the form of a decision
tree and obtain the optimal decision.
 Redraw further the decision tree by introducing the
following information:
a) If it manufactures itself and sales are medium or high,
then company has the opportunity of developing a new
version of its telephone.
b) From the past experience, company estimates that there is
a 50% chance of successful development.
c) The cost of development is Rs 1,50,000 and returns after
deduction of development cost are Rs 3,00,000 and Rs
1,00,000 for high and medium sales respectively.
EXAMPLE XVI SOLUTION
High Sales (0.20) (Event) 800
Medium Sales (0.30) (Event)

Manufacture
A 300
(Action)
Low Sales (0.50) (Event)

-100
High Sales (0.20) (Event)
180

1 B 180
Sell all rights
(Action)
180

350
Royalties
(Action)

C 200

100
EXAMPLE XVI SOLUTION
High Sales (0.20) (Event) 800

Manufacture
200 300
(Action)

-100
180

200 180 180
Sell all rights
(Action)
180

350
Royalties
(Action)

180 200

100
Success (0.50)
(Event) 300
Develop (Action) A
High Sales Failure
(0.20) (Event) 1 (0.50) (Event) -150
Do not
Develop (Action)
0
Manufacture
C -100
(Action) Do not
Develop (Action) 0
Success (0.50)
Medium Sales
2 (Event) 100
(0.30) (Event)
Develop
(Action) B
180 Failure
High Sales (0.20) (Event) (0.50) (Event) -150
3 D 180
Sell all rights Low Sales (0.50) (Event)
(Action) 180
Royalties
(Action) High Sales (0.20) (Event) 350

E 200
100
Success (0.50)
(Event) 300
(800 + 75) Develop (Action) 75
High Sales Failure
(0.20) (Event) 875 (0.50) (Event) -150
Do not
Develop (Action)
0
Manufacture
215 -100
(Action) Do not
(300 + 0) Develop (Action) 0
Success (0.50)
Medium Sales
300 (Event) 100
(0.30) (Event)
Develop
(Action) -25
180 Failure
High Sales (0.20) (Event) (0.50) (Event) -150
215 180 180
Sell all rights Low Sales (0.50) (Event)
(Action) 180
Royalties
(Action) High Sales (0.20) (Event) 350

180 200
100
EXAMPLE NO. 39 SOLUTION
12000
Continue with
Current Process
(Action) Success (0.90)
(Event)
16000
Conduct
Research R1 A
(Action)
Failure
(0.10) (Event) -10000
1 Success (0.60) 18000
(Event)
Conduct
B
Research R2
(Action) Failure
(0.40) (Event)
-6000
Pay Royalty
(Action)
15000
120000
Success (x)
(Event)
Start New
A
Business Venture
(Action)
Failure (1-x)
(Event)
1 -20000
Put in Bank
(Action)
60000
120000
Success (x)
(Event)
Start New
A
Business Venture
(Action)
Failure (1-x)
(Event)
1 -20000
Put in Bank
(Action)
60000
60000 = 120000x + [-20000 (1 - x)]

60000 = 120000x – 20000 + 20000x
80000 = 140000x
x = 0.5714 & (1 - x) = 0.4286
12000
Continue with
Current Process
(Action) Success (0.90)
(Event)
16000
Conduct
Research R1 13400
(Action)
Failure
(0.10) (Event) -10000
15000 Success (0.60) 18000
(Event)
Conduct 8400
Research R2 Failure
(Action) (0.40) (Event)
-6000
Pay Royalty
(Action)
15000
EXAMPLE NO. 40 SOLUTION Success (0.55)
(Event) 19
A Design C
Design A
Failure (Action)
(Action)
(0.45) (Event) 1 12.5
(12.5 - 6)
Design C
(Action)
3 12.5
Win (0.60)
(Event)
C Design B
Bid (Action) Success (0.50)
(Action) (Event) 16
Lose (0.40)
(Event) B Design C
(Action)
4 -5 Failure
2 12.5
(0.50) (Event)
Do not Bid
(Action)
0
EXAMPLE NO. 40 SOLUTION Success (0.55)
(Event) 19
13.375 Design C
Failure (Action)
Design A (0.45) (Event) 6.5 12.5
(Action)
(12.5 - 6)
(13.375 - 5) Design C
(Action)
8.375 12.5
Win (0.60)
(Event)
3.025 Design B
Success (0.70)
Bid
(Action)
(Event) 16
(Action)
Lose (0.40) 12.25 Design C
(Event) (Action)
3.025
Failure
(0.30) (Event) 3.5 12.5
-5
(12.5 - 9)
Do not Bid
(Action)
0
THANKS

Ch. 1 Decision Theory

Uploaded by

Copyright:

Available Formats

Ch. 1 Decision Theory

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ch. 1 Decision Theory

Uploaded by

Copyright:

Available Formats

UNIT 1

DECISION THEORY &

 Introduction to Decision-Making Process (0.5 Hours)

 Steps of Decision-Making Process (0.5 Hours)

 Types of Decision Making Environments (0.5 Hours)

 Decision Making under Uncertainty (1.5 Hours)

 Decision Making under Risk (3 Hours)

 Decision-Tree Analysis (1.5 Hours)

 To arrive at such a decision,

A decision may be defined as the

These are the finite number of options

Number and type of alternatives depend

The decision alternatives may be

It is an event or scenario that is not under

It may be a state of economy, a weather

States of nature are mutually exclusive

Itis the numerical value (outcome which is

The payoffs are measured within a specified

2. List all possible

4. Express the payoffs (Pij) resulting from

5. Apply an appropriate model to select

The degree of Knowledge is divided into 4

2. Decision Making under Risk

3. Decision Making under

The Likely demand (units) of the products is given below:-

If the sale price of each type of product is Rs 25, then

The payoff in this case is determined by

So, the payoff for each pair of demand (state

D1B = (3000 x 25) – 35000 – (3000 x 9)

A 14000 66000 118000

B 13000 77000 141000

C 1000 73000 145000

When the probability of

1. Optimism (Maximax or minimin) Criterion

2. Pessimism (Maximin or minimax) Criterion

3. Equal Probabilities (Laplace) Criterion

4. Coefficient of Optimism (Hurwicz) Criterion

5. Regret (Savage) or Opportunity loss Criterion

a) Find the maximum (or

c) The action corresponding to this

a) Find the minimum (or

c) The action corresponding to

a) Calculate the average of all the

Where, α = Coefficient of Optimism

a) Prepare a new regret table by

c) Select the minimum value from

d) The action corresponding to this

High Moderate Low Nil

Expand 50000 25000 -25000 -45000

Construct 70000 30000 -40000 -80000

Subcontract 30000 15000 -1000 -10000

1. Optimism (Maximax or minimin) Criterion

2. Pessimism (Maximin or minimax) Criterion

3. Equal Probabilities (Laplace) Criterion

4. Coefficient of Optimism (Hurwicz) Criterion

5. Regret (Savage) or Opportunity loss Criterion

Cost (loss) Matrix

1. Optimism (Maximax or minimin) Criterion

2. Pessimism (Maximin or minimax) Criterion