DECISION ANALYSIS

1. The payoffs (in ₹) of three acts A1, A2 and A3 and the possible states of nature S1, S2 and
S3 are given in the adjoining table.

States of nature Acts

A1 A2 A3
S1 -20 -50 200
S2 200 -100 -50
S3 400 600 300
The probabilities of the states of nature are 0.3, 0.4 and 0.3 respectively.
Determine the optimal act using the expectation principle.
2. Marketing staff of a certain industrial organization has submitted the following payoff table,
giving profits in million rupees, concerning a certain proposal depending upon the rate of
technological advance in the next three years:

Technological Decision
Advance Accept Reject
Much 2 3
Little 5 2
None -1 4
The probabilities are 0.2, 0.5 and 0.3 for Much, Little and None technological advance
respectively. What decision should be taken?
3. A person wants to invest in one of the three alternative investment plans: Stock, Bonds,
Debentures. It is assumed that the person wishes to invest all of the funds in a plan. The
payoff matrix based on three potential economic conditions is given in the adjoining table:

Alternative Economic Conditions

Investment High Growth (₹) Normal Growth (₹) Slow Growth (₹)
Stock 10,000 7,000 3,000
Bonds 8,000 6,000 1,000
Debentures 6,000 6,000 6,000
Determine the best investment plan using each of the following criteria:
(i) Laplace, (ii) Maximin, (iii) Maximax
4. Given is the following payoff matrix

States of Nature Probability Courses of Action

Do not expand Expand 200 Expand 400
units units
High Demand 0.4 2,500 3,500 5,000
Medium Demand 0.4 2,500 3,500 2,500
Low Demand 0.2 2,500 1,500 1,000
What should be the decision if we use: (i) EMV criterion, (ii) The Maximin criterion, (iii)
The maximax criterion, (iv) Minimax regret criterion?
5. The research director of XYZ Pharmaceutical Laboratory has to decide about one of three
influenza vaccines (P1, P2, P3) which should be funded for mass production. Payoffs depend
upon the type of influenza outbreak (S1, S2, S3, S4) that is most persuasive in the next year.
The payoff matrix, with profits (in millions of rupees), is given below:

States of Nature Courses of Action

P1 P2 P3
S1 10 8 -15
S2 4 12 12
S3 0 -5 8
S4 -2 -10 8
Prior to acquiring any additional information about the occurrence of states of nature, the
director’s probability judgements are: P(S1) = 0.2 and P(S2) = 0.2, P(S3) = 0.5 and P(S4) =
0.1.
(i) If the director could consult an authority who could tell him which state will occur,
what is the expected value of this information using above payoff matrix.
(ii) Verify your answer by calculating EVPI from the loss matrix.
6. An executive has to make a decision. He has four alternatives D1, D2, D3 and D4. When the
decision has been made events may lead such that any of the four results may occur. The
results are R1, R2, R3 and R4. Probabilities of occurrence of these results are as follows:
R1 = 0.5, R2 = 0.2, R3 = 0.2 and R4 = 0.1
The matrix of payoff between the decision and the results is indicated in the adjoining table:

R1 R2 R3 R4
D1 14 9 10 5
D2 11 10 8 7
D3 9 10 10 11
D4 8 10 11 13
Show this decision situation in the form of a decision tree and indicate the most preferred
decision and corresponding expected value.
7. 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 metres, some
persons in that area drilled it further up to 25 metres but only 20 % struck water at that level.
The prevailing cost of drilling is ₹ 500/m. The Finance Manager estimated that in case he
does not get water in his own well, he will have to pay ₹ 15,000 to buy water from outside
for the same period of getting water from the well. The following decisions are considered:
(i) Do not drill any well;
(ii) Drill upto 20 m, and
(iii) If no water is found at 20 metres, drill further up to 25 m.

Draw an appropriate decision tree and determine the Finance Manager’s optimal strategy.
8. Expected return (in million rupees) from the sale of three machines A, B and C under
expected market condition as poor (S1), Fair (S2) and Good (S3) are given in the following
table below:

Sales Courses of Action

Poor (S1) Fair (S2) Good (S3)
S1 0.5 1.0 1.5
S2 0 1.5 2.5
S3 -1.5 0.5 3.5
Chance of market at states S1, S2 and S3 are 30%, 50% and 20% respectively. But the
market research finds the actual chances of states of market as follows:

Actual State M1 (Poor) M2 (Fair) M3 (Good)

S1 0.7 0.2 0.1
S2 0.2 0.7 0.1
S3 0 0.2 0.8
Find (i) Conditional expected loss table
(ii) Expected Value of Perfect Information (EVPI).
(iii) Expected loss table on the basis of the results of market research.

Illustration
Suppose a electrical good has a resource base to buy for resale purposes in a market, electric
irons in the range of 0 to 4. His resource base permits him to buy nothing or 1 or 2 or 3 or 4
units. These are his alternative courses of action or strategies. The demand for electric irons in
any month is something beyond his control and hence is a state of nature. Let us presume that the
dealer does not know how many units will be bought from him by the customers. The demand
could be anything from 0 to 4. The dealer can buy each unit of electric iron @ ₹ 40 and sell it at
₹ 45 each, his margin being ₹ 5 per unit. Assume the stock on hand is valueless. Portray in a
payoff table the EMV.

COMPUTATION OF EXPECTED MONETRAY VALUE (EMV)

States Probabilit Conditional Payoff (₹) Expected Payoff (₹)

of y Courses of action Courses of action
natur A1(0 A2(1 A3(2 A4(3 A5(4 A1(0 A2(1 A3(2 A4(3 A5(4
e ) ) ) ) )
) ) ) ) )
(1) (2) (3) (4) (5) (6)
(1) x (1) x (1) x (1) x (1) x
(2) (3) (4) (5) (6)
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 0.30 -2.1 -4.5 -6.9
S3(2) 0.20 0 5 10 -30 -70 0 1.0 2.0 -6.0 -14.0
S4(3) 0.30 0 5 10 15 -25 0 1.5 3.0 4.5 -7.5
S5(4) 0.40 0 5 10 15 20 0 2.0 4.0 6.0 8.0
EMV 0 3.2 3.7 -4.8 -26.8
Conditional payoff value = (Marginal profit (Units sold) – (Marginal Loss) (Units not sold)
= (₹ 45 - ₹ 40) (Units sold) – (₹ 40)(Units not sold)
PAYOFF AND REGRET TABLE
States of Conditional Payoff (₹) Conditional Opportunity Loss (₹)
nature Courses of action (Strategies Courses of action (Strategies Possible Supply)
(Probable Possible Supply)
Demand) 0 1 2 3 4 0 1 2 3 4
0 0 -40 -80 -120 -160 0–0=0 0 – (-40) = 40 0 – (-80) = 80 0 – (-120) = 120 0 – (-160) = 160
1 0 5 -35 -75 -115 5–0=5 5–5=0 5 – (-35) = 40 5 – (-75) = 80 5 – (-115) = 120
2 0 5 10 -30 -70 10 – 0 = 10 10 – 5 = 5 10 – 10 = 0 10 – (-30) = 40 10 – (-70) = 80
3 0 5 10 15 -25 15 – 0 = 15 15 – 5 = 10 15 – 10 = 5 15 – 15 = 0 15 – (-25) = 40
4 0 5 10 15 20 20 – 0 = 20 20 – 5 = 15 20 – 10 = 10 20 – 15 = 5 20 – 20 = 0

States of Probability Conditional Opportunity Loss (₹) Expected Opportunity Loss (₹)
nature Courses of action
0 1 2 3 4 0 1 2 3 4
0 0.04 0 40 80 120 160 0 1.6 3.2 4.8 6.4
1 0.06 5 0 40 80 120 0.3 0 2.4 4.8 7.2
2 0.20 10 5 0 40 80 2 1 0 8 16
3 0.30 15 10 5 0 40 4.5 3 1.5 0 12
4 0.40 20 15 10 5 0 8 6 4 2 0
Expected Opportunity Loss (EOL) 14.8 11.6 11.1 19.6 41.6

Solutions
1. COMPUTATION OF EXPECTED MONETRAY VALUE (EMV)

States Probability Conditional Payoff(₹) Expected Payoff (₹)

of P(S) Acts Acts
 Nature
(Sj)
A1 A2 A3 A1 A2 A3
S1 0.3 -20 -50 200 -6 -15 60
S2 0.4 200 -100 -50 80 -40 -20
S3 0.3 400 600 300 120 180 90
Expected Monetary Value (EMV) 194 125 130
The maximum value of EMV is corresponding to act A1. Hence, according to the EMV
criterion, the optimal act is A1.

2. COMPUTATION OF EMV FOR VARIOUS ACTS

Technological Probability Conditional Payoff Expected Payoff

Advance Accepting Rejecting Accepting Rejecting
Much 0.2 2 3 0.4 0.6
Little 0.5 5 2 2.5 1.0
None 0.3 -1 4 -0.3 1.2
Expected Monetary Value (EMV) 2.6 2.8
Since EMV of rejecting the proposal is 2.8 which is more than EMV of accepting the
proposal, the decision should be ‘reject the proposal’.

3. Let HG: High Growth, NG: Normal Growth, SG: Slow Growth
PAYOFF TABLE (in Rupees)

Act States of nature Row Row Row

(Investment S1: S2: NG S3: SG Minimum Maximum Total
HG
(1) (2) (3) (4) (5) (6) (7)
A1: Stocks 10,000 7,000 3,000 3,000 10,000 20,000
A2: Bonds 8,000 6,000 1,000 1,000 8,000 15,000
A3: Debentures 6,000 6,000 6,000 6,000 6,000 18,000
Probability 1/3 1/3 1/3 Column (5) Column (6)
Max. = 6,000 Max. = 10,000
(i) Laplace Criterion
EMV (A1: Stocks) = ₹ 1/3(10,000 + 7,000 + 3,000) = ₹ 20,000/3 = ₹ 6,666.67
EMV (A2: Bonds) = ₹ 1/3(8,000 + 6,000 + 1,000) = ₹ 15,000/3 = ₹ 5,000
EMV (A3: Debentures) = ₹ 1/3(6,000 + 6,000 + 6,000) = ₹ 18,000/3 = ₹ 6,000
Max. (EMV) = ₹ 6,666.67 which corresponds to acts A1. Hence, under Laplace criterion act
A1: Stock, can be taken as the optimal act.
(ii) Maximin Criterion
From column (5) of the above Table, we get
Maximum (Minimum Payoffs) = ₹ 6,000, which corresponds to act A3.
Hence, under the Maximin criterion, act A3: Debenture is the optimal choice
 (iii) Maximax Criterion
From column (6) of the above Table, we get
Maximum (Maximum Payoffs) = ₹ 10,000, which corresponds to act A1.
Hence, under the Maximax criterion, act A1: Stock is the optimal choice
4. Payoff Table

Act Probability Conditional Payoff (₹) Expected Payoff (₹)

(Investment Courses of action Courses of action
A1: Do A2: A3: A1: Do A2: A3:
not Expand Expand not Expand Expand
expand 200 units 400 units expand 200 units 400 units
(1) (2) (3) (4) (1) x (2) (1) x (3) (1) x (4)
S1: High 0.4 2,500 3,500 5,000 1,000 1,400 2,000
Demand
S2: Medium 0.4 2,500 3,500 2,500 1,000 1,400 1,000
Demand
S3: Low 0.2 2,500 1,500 1,000 500 300 200
Demand
EMV 2,500 3,100 3,200
Minimum Payoff (₹) 2,500 1,500 1,000
Maximum Payoff (₹) 2,500 3,500 5,000
(i) EMV criterion thus suggests that we should decide to expand 400 units since EMV
3,200 is highest.
(ii) In the maximin criterion the strategy for which minimum payoff is maximum is
chosen. The minimum payoff values corresponding to the strategies: Do not expand,
Expand 200 units, and Expand 400 units, are 2,500; 1,500 and 1,000 respectively. Of
these payoffs 2,500 is maximum which corresponds to the strategy ‘Do not expand’.
Therefore, a decision maker using Maximin criterion would decide ‘Not to expand’.

Overall maximum payoff values (due to high demand) are ₹ 5,000 that corresponds
to the act – Expand 400 units. By using maximax criterion the decision maker would
decide ‘Expanding 400 units’.

Minimax Regret:
In this criterion profits are transformed into opportunity losses (or regret). A regret
matrix is obtained from the payoff matrix by subtracting each of the values in a row
from the largest payoff value in the row. Under this approach the decision – maker
identifies the maximum regret for each act and selects the act due to which maximum
regret value is minimum. This may be achieved by selecting the act which maximum
regret (i.e. column maximum of the regret matrix) is minimum.

Regret matrix of the previous payoff matrix is as follows:

States of Nature Probabilit Courses of Action (Possible Supply)

y A1: Do not A2: Expand A3: Expand 400
 Expand 200 units units
S1: High Demand 0.4 5,000 – 2,500 = 5,000 – 3,500 = 5,000 – 5,000 = 0
2,500 1,500
S2: Medium 0.4 3,500 – 2,500 = 3,500 – 3,500 = 3,500 – 2,500 =
Demand 1,000 0 1,000
S3: Low Demand 0.2 2,500 – 2,500 = 0 2,500 – 1,500 = 2,500 – 1,000 =
1,000 1,500
Maximum Regret 2,500 1,500 1,500
The decision – maker must choose ‘Expand 200 units’ or ‘Expand 400 units’ for it
minimizes the maximum possible return.
5. COMPUTATION OF EXPECTED PAYOFF

Act Probability Conditional Payoff (₹) Expected Payoff (₹)

(Investmen Courses of action Courses of action
t P1 P2 P3 P1 P2 P3
(1) (2) (3) (4) (1) x (2) (1) x (3) (1) x (4)
S1 0.2 10 8 -15 2.0 1.6 -3.0
S2 0.2 4 12 12 0.8 2.4 2.4
S3 0.5 0 -5 8 0 -2.5 4.0
S4 0.1 -2 -10 8 -0.2 -1.0 0.8
EMV 2.6 0.5 4.2
From the table, we find that the highest prior expected value is 4.2 (million rupees). Prior
expected value of selecting the optimal act after learning which state will occur
= 10 x 0.2 + 12 x 0.2 + 8 x 0.5 + 8 x 0.1 = 9.2
Expected value of perfect information = 9.2 – 4.2 = ₹ 5 million
6. Decision Tree Diagram
R1

D1 R2

R3

D2 R4

D3

D4
 Monetary Value Prob. Expected Value EMV
14 (R1) 0.5 7.0
D1 9 (R2) 0.2 1.8 11.3
10 (R3) 0.2 2.0
5 (R4) 0.1 0.5
11 0.5 5.5
D2 10 0.2 2.0 9.8
8 0.2 1.6
7 0.1 0.7
9 0.5 4.5
D3 10 0.2 2.0 9.6
10 0.2 2.0
11 0.1 1.1
8 0.5 4.0
D4 10 0.2 2.0 9.5
11 0.2 2.2
13 0.1 1.3
The most preferred decision at the decision node 1 is found by calculating expected value of each decision branch
and selecting the path (course of action) with high value.

Since node D1 has the highest EMV, the decision at node A will be choose the course of action D1.

7. Based on the given information, the decision tree is shown in the figure below
Do not drill
₹ 15,000 No water (0.8) ₹ 15,000 +
500
X 25 = 27,500
Drill upto 25m

D2 No water (0.3) Water (0.2) 500 x 25

D1 = ₹ 12,500

Stop ₹ 15,000 + 500 x 20

= ₹ 25,000
Drill upto 20 m

Water (0.7) ₹ 500 x 20 m

= ₹ 10,000

Decision Options Expected Cost Decision

Node
1 Drill upto 25m 0.8 x ₹ 27,500 + 0.2 x ₹ 12,500 = ₹ Drill upto 25 m
24,500
Stop
₹ 25,000
2 Do not drill ₹ 15,000 Drill upto 20 m

Drill upto 20 m 0.3 x 24,500 + 0.7 x 10,000 = ₹ 14,350

From the analysis table, it may be observed that decision at node 2 implies that if it is decided to drill up to 20 m and
water is not found, then drilling up to 25 m should be done. At node 1, the decision taken is to drill up to 20 m as it
involved lower expected cost. Thus, the optimal strategy is to drill up to 20 m and if water is not struck then drill
further to 25 m.

States of Nature Probabilit Courses of Action (Possible Supply)

y A1: Do not A2: Expand A3: Expand 400
Expand 200 units units
S1: High Demand 0.4 5,000 – 2,500 = 5,000 – 3,500 = 5,000 – 5,000 = 0
2,500 1,500
S2: Medium 0.4 3,500 – 2,500 = 3,500 – 3,500 = 3,500 – 2,500 =
Demand 1,000 0 1,000
S3: Low Demand 0.2 2,500 – 2,500 = 0 2,500 – 1,500 = 2,500 – 1,000 =
1,000 1,500
Maximum Regret 2,500 1,500 1,500
8. (i) (a) The conditional profit table is given below:

States of Prior Courses of Action (Buying Decision)

Nature Probability A B C
S1: Poor 0.30 0.5 0 -1.5
S2: Fair 0.50 1.0 1.5 0.5
S3: Good 0.20 1.5 2.5 3.5
(b) Subtracting the payoffs against each event from the largest payoffs (market*) gives the conditional
opportunity losses (COL) as shown in the table below:

Act Probability Conditional Loss (₹) Expected Opportunity Loss (₹)

(Investmen Courses of action Courses of action
t A B C A B C
S1 0.30 0 0.5 2.0 0 0.15 0.60
S2 0.50 0.5 0 1.0 0.25 0 0.50
S3 0.20 2.0 1.0 0 0.40 0.20 0
EMV 0.65 0.35 1.10
(ii) EOL for machine B is least (0.35). Under perfect information, the opportunity loss would be zero, so
the expected value under EVPI is 0.35.
(iii) The margin and joint prob. Is computed as under:

Act Probability Conditional Prob. Joint Prob.

(Investmen Courses of action Courses of action
t
S1 0.30 0.7 0.2 0.1 0.21 0.06 0.03
S2 0.50 0.2 0.7 0.1 0.10 0.35 0.05
S3 0.20 0 0.2 0.8 0 0.04 0.16
Total P(M1) P(M2) = P(M3) =
= 0.31 0.45 0.24
Revising the prior prob with the help of Bayes’ Theorem, the reqd. posterior prob. Are computed as
below:

Outcome Prob. States of nature Posterior prob.

M1 0.31 S1 0.21/0.31 = 0.677
S2 0.10/0.31 = 0.323
S3 0/0.31 = 0
M2 0.45 S1 0.06/0.45 = 0.133
S2 0.35/0.45 = 0.778
S3 0.04/0.45 = 0.089
M3 0.24 S1 0.03/0.24 = 0.125
S2 0.05/0.24 = 0.208
S3 0.16/0.24 = 0.667

Act I II
(Investment Prob COL EOL A B C
S1 0 0.5 2.0 0 0.15 0.60
S2 0.5 0 1.0 0.25 0 0.50
S3 2.0 1.0 0 0.40 0.20 0