Exercise-9-

Given the following Payoff tables : where the table represents profits :

State which can be chosen as the best act using:

(a) Maximax, (b) Maximin, (c) Minimax regret (savage criterion),
(d) Equal likelihood (Laplace),
(e) Hurwicz Alpha criterion α=0.4

Decision making under uncertainty.

a) Maximax :
Max( Max Ai) = Max ( 8, 6, 6 ) = 8
Decision : Select A1

b) Maxmin :
Max(Min Ai) =Max (-1, 0 ,0) =0
Decision : Select A2 or A3

c) Minimax :
Loss table

Alternatives States of nature

S1 S2 S3 S4
A1 3 0 0 5
A2 0 0 6 4
A3 6 0 2 0
Min(Max Ai)= Min(5, 6, 6) =5
Decision : Select A1

d) Laplace :
we associate equal probability for event say say ¼
Expected pay-offs are :
𝟑+𝟓+𝟖−𝟏 𝟏𝟓
E(A1)= = = 𝟑. 𝟕𝟓
𝟒 𝟒
𝟔+𝟓+𝟐+𝟎 𝟏𝟑
E(A2)= = = 𝟑. 𝟐𝟓
𝟒 𝟒
𝟎+𝟓+𝟔+𝟒 𝟏𝟓
E(A3)= = = 𝟑. 𝟕𝟓
𝟒 𝟒
A1 and A3 has the maximum expected pay-off
Decision : Select A1 or A3

EX. 9 OP122 1
 e) Hurwicz :

D= α (Maximum in Ai) + (1 – α )(Minimum in Ai)

D(A1)=(0.4*8)+(0.6*-1) =2.6
D(A2)=(0.4*8)+(0.6*0) =2.4
D(A3)=(0.4*6)+(0.6*0) =2.4

D(A1) has the maximum

Decision : Select A1

Given : P(S1)=0.6 , P(S2)=0.1 , P(S3)=0.2 , P(S4)=0.1 in table (A) ,Find:

(f)Expected Monetary Value (EMV) ,
(g)Expected Opportunity Loss (EOL) ,
(k)Expected Value of Perfect Information (EVPI)

Alternatives States of nature

S1 S2 S3 S4
A1 3 5 8 -1
A2 6 5 2 0
A3 0 5 6 4
probability 0.6 0.1 0.2 0.1

Decision making under Risk.

f) Expected Monetary Value (EMV)
EMV for A1 = (3*0.6)+(5*0.1)+(8*0.2)+(-1*0.1)= 3.8
EMV for A2 = (6*0.6)+(5*0.1)+(2*0.2)+(0*0.1)= 4.5
EMV for A3 = (0*0.6)+(5*0.1)+(6*0.2)+(4*0.1)= 2.1

EMV for A2 is greater

Decision : Select A2

EX. 9 OP122 2
 g) Expected Opportunity Loss (EOL) ,
Loss table

Alternatives States of nature

S1 S2 S3 S4
A1 3 0 0 5
A2 0 0 6 4
A3 6 0 2 0
probability 0.6 0.1 0.2 0.1

ELO for A1 = (3*0.6)+(0*0.1)+(0*0.2)+(5*0.1)= 2.3

ELO for A2 = (0*0.6)+(0*0.1)+(6*0.2)+(4*0.1)= 1.6
ELO for A3 = (6*0.6)+(0*0.1)+(2*0.2)+(0*0.1)= 4

EMV for A2 is least

Decision : Select A2

h) Expected Value of Perfect Information (EVPI)

Max. profit of probability Expected value
each state (S) (= prob. * profit)
S1 6 0.6 3.6
S2 5 0.1 0.5
S3 8 0.2 1.6
S4 4 0.1 0.4
Sum 6.1
𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑝𝑟𝑜𝑓𝑖𝑡 𝑤𝑖𝑡ℎ 𝑝𝑒𝑟𝑓𝑒𝑐𝑡
𝑖𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛

𝐸𝑉𝑃𝐼 = 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑝𝑟𝑜𝑓𝑖𝑡 𝑤𝑖𝑡ℎ 𝑝𝑒𝑟𝑓𝑒𝑐𝑡 𝑖𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 − max 𝐸𝑀𝑉

= 6.1 – 4.5
= 1.6

EX. 9 OP122 3
(H.W)
Example: An agricultural company wants to decide which commodity should stock to get
maximum profit. It was supplied with the following information. The probability that the
monsoon will be excess, normal and deficient is 0.40,0.30 and 0.30 . The estimated profit or
loss three commodities in respect of these different kinds of monsoon are:

Profit per 1 ton

Monsoon Excess Normal Deficient
Rice 10,000 -4,000 15,000
Wheat 4,000 -3,000 8,000
Maize 4,000 1,000 -1,000

Determine the optimal decision under each of the following decision criteria and show how you
arrived at it:

(a) Maximax, (b) Maximin, (c) Minimax regret (savage criterion), (d) Equal likelihood
(Laplace), (e) Hurwicz Alpha criterion α=0.8, (f)EMV, (g)EOL, (k)EVPI .

 Decision making under uncertainty.

1. Maximax criterion (OPTIMISM)
Alternative Rice Wheat Maize
Monsoon
Excess 10,000 4,000 4,000
Normal -4,000 -3,000 1,000
Deficient 15,000 8,000 -1,000
MAX 15,000 8,000 4,000
The maximum of column maxima is 15,000. Hence, the company should adopt Rice commodity.

2. Maximin criterion (pessimism).

Alternative Rice Wheat Maize

Monsoon
Excess 10,000 4,000 4,000
Normal -4,000 -3,000 1,000
Deficient 15,000 8,000 -1,000
Min -4,000 -3,000 -1,000
The maximum of column minima is -1,000. Hence, the company should adopt Maize commodity.
3. Minimax regret.
Note: From the given payoff matrix,
Alternative Rice Wheat Maize develop an opportunity-loss (or regret)
Monsoon matrix as follows:
(i) Find the best payoff corresponding
Excess 10,000 4,000 4,000 to each state of nature
Normal -4,000 -3,000 1,000 (ii) Subtract all other payoff values in
that row from this value.
Deficient 15,000 8,000 -1,000

EX. 9 OP122 4
 Opportunity Loss Table
Alternative Rice Wheat Maize
Monsoon
Excess 10,000-10,000= 10,000-4,000= 10,000-4,000=
0 6,000 6,000
Normal 1,000-(-4,000)= 1,000-(-3,000) = 1,000-1,000=
5,000 4,000 0
Deficient 15,000-15,000= 15,000-8,000 = 15,000-(-1,000)=
0 7,000 16,000
MAX 5,000 7,000 16,000
Decision: Rice

4. Equal likelihood (Laplace).

Alternative Rice Wheat Maize

Monsoon
Excess 10,000 4,000 4,000
Normal -4,000 -3,000 1,000
Deficient 15,000 8,000 -1,000
1 1 1
(10,000 (4,000 − 3,000 (4,000 − 1,000
Expected 3 3 3
− 4,000 + 8,000) + 1,000)
(average) + 15,000) = 3000 = 1333.33
= 7000
Decision: Rice

5. Hurwicz Alpha criterion α=0.8.

Alternative Rice Wheat Maize

Monsoon
Excess 10,000 4,000 4,000
Normal -4,000 -3,000 1,000
Deficient 15,000 8,000 -1,000
weighted 0.8(15,000)+0.2(- 0.8(8,0000+0.2(- 0.8*4,000+0.2(-
average: Di 4,000)= 11,200 3,000)= 5,800 1,000)= 3000

D= α (Maximum in column) + (1 – α )(Minimum in column)

Decision: Rice

EX. 9 OP122 5
  Decision making under Risk.

1. Expected monetary value (EMV).

Alternative Rice Wheat Maize probability
Monsoon
Excess 10,000 4,000 4,000 0.4
Normal -4,000 -3,000 1,000 0.30
Deficient 15,000 8,000 -1,000 0.30
EMV 0.4(10,000)+0.3(- 0.4(4,000)+0.3(- 0.4(4,000)+0.3(-
4,000)+0.3(15,000)= 3,000)+0.3(8,000)= 1,000)+0.3(1,000)=
7300 3100 1600
Decision: Rice

2. Expected Opportunity Loss (EOL).

Opportunity Loss Table

Rice Wheat Maize probability

Alternative
Monsoon
Excess 0 6,000 6,000 0.4
Normal 5,000 4,000 0 0.30
Deficient 0 7,000 16,000 0.30
EOL 0.3(5,000) 0.4(6,000)+0.3(4,000)+0.3(7,000) 0.4(6,000)+0.3(16,000)
=1,500 =5,700 =7,200
Decision: Rice

3. Expected Value of Perfect Information (EVPI).

Alternative Rice Wheat Maize probability

Monsoon
Excess 10,000 4,000 4,000 0.4
Normal -4,000 -3,000 1,000 0.30
Deficient 15,000 8,000 -1,000 0.30
EMV 0.4(10,000)+0.3(- 0.4(4,000)+0.3(- 0.4(4,000)+0.3(-
4,000)+0.3(15,000)= 3,000)+0.3(8,000)= 1,000)+0.3(1,000)=
7300 3100 1600

𝐸𝑉𝑃𝐼 = 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑝𝑟𝑜𝑓𝑖𝑡 𝑤𝑖𝑡ℎ 𝑝𝑒𝑟𝑓𝑒𝑐𝑡 𝑖𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 − max 𝐸𝑀𝑉

= ∑𝑚
𝑖=1 𝑝𝑖 max(𝑟𝑖𝑗 ) − max 𝐸𝑀𝑉.
𝐸𝑉𝑃𝐼 = 10,000(0.4) + 0.3(1,000) + 0.3(15,000) − 7,300 = 8,800 − 7,300 = 1,500

EX. 9 OP122 6