You are on page 1of 18

DECISION THEORY

Sensitivity Analysis

Adelberto Ll. Areola


October 9, 2019
Given this payoff table:
Alternative SON 1 SON 2
A $ 120 20
B 60 40
C 10 110
D 90 90

A. Determine the range of P(1) for which


each alternative would be best, treating
the payoffs as profits.
B. Answer part A, treating the payoffs as
costs.
Solution for part A
(payoffs as profits)
Step 1. Prepare the coordinate axes.

SON 2 SON 1

P(1)
0 1
Step 2. Graph the line corresponding to each
alternative decision.
Alternative SON 1 SON 2 SON 2 SON 1
120
A $ 120 20
110
B 60 40

C 10 110
90 90
D 90 90

60

40

20

10

P(1)
0 1
Step 3. Identify the optimal (maximum) profit
boundaries for the range 0 ≤ P(1) ≤ 1.
Alternative SON 1 SON 2 SON 2 SON 1
120
A $ 120 20
110
B 60 40 C A
C 10 110 D
90 90
D 90 90

60

40

20

10

P(1)
0 1
Step 4. Identify the optimal profit points (OPPs).

Alternative SON 1 SON 2 SON 2 SON 1


120
A $ 120 20
110
B 60 40 C A
C 10 110 D OPP 2
90 90
D 90 90 OPP 1

60

40

20

10

P(1)
0 1
Step 5. Find the equation of the lines that
intersect through each OPP.
Alternative SON 1 SON 2 SON 2 SON 1
120
A $ 120 20
110
B 60 40 C A
C 10 110 D OPP 2
90 90
D 90 90 OPP 1

OPP 1 60
C: y = 110 – 100x
D: y = 90 40

OPP 2
20
A: y = 20 + 100x
10
D: y = 90
P(1)
0 1
Step 6. Find the x-coordinate of each OPP by
equating the intersecting lines and solving for x.
SON 2 SON 1
OPP 1 120
110 – 100x = 90 110
C A
– 100x = 90 – 110
D OPP 2
– 100x = – 20 90 90
OPP 1
x = 0.2

OPP 2 60

20 + 100x = 90
100x = 90 – 20 40

100x = 70
x = 0.7 20

10

P(1)
0 1
Step 7. State your decision to maximize profit.

SON 2 SON 1
A: 0.7 < P(1) ≤ 1.0 120
B: Never optimal 110
C A
C: 0 ≤ P(1) < 0.2
D OPP 2
D: 0.2 < P(1) < 0.7 90 90
OPP 1

If probability of SON 1 is 60

less than 0.2, choose C;


if between 0.2 and 0.7, 40

choose D;
if greater than 0.7, 20

choose A. 10

P(1)
0 1
Solution for part B
(payoffs as costs)
Step 1. Prepare the coordinate axes.

SON 2 SON 1

P(1)
0 1
Step 2. Graph the line corresponding to each
alternative decision.
Alternative SON 1 SON 2 SON 2 SON 1
120
A $ 120 20
110
B 60 40

C 10 110
90 90
D 90 90

60

40

20

10

P(1)
0 1
Step 3. Identify the optimal (minimum) cost
boundaries for the range 0 ≤ P(1) ≤ 1.
Alternative SON 1 SON 2 SON 2 SON 1
120
A $ 120 20
110
B 60 40

C 10 110
90 90
D 90 90

60

40 B

A
20 C
10

P(1)
0 1
Step 4. Identify the optimal cost points (OCPs).

Alternative SON 1 SON 2 SON 2 SON 1


120
A $ 120 20
110
B 60 40

C 10 110
90 90
D 90 90

60

OCP 1
B OCP 2
40

A
20 C
10

P(1)
0 1
Step 5. Find the equation of the lines that
intersect through each OCP.
Alternative SON 1 SON 2 SON 2 SON 1
120
A $ 120 20
110
B 60 40

C 10 110
90 90
D 90 90

OCP 1 60
A: y = 20 + 100x OCP 1
B: y = 40 + 20x B OCP 2
40

A
OCP 2 C
20
B: y = 40 + 20x
10
C: y = 110 – 100x
P(1)
0 1
Step 6. Find the x-coordinate of each OCP by
equating the intersecting lines and solving for x.
SON 2 SON 1
OCP 1 120
20 + 100x = 40 + 20x 110

100x – 20x = 40 – 20
80x = 20 90 90
x = 0.25

OCP 2 60

40 + 20x = 110 – 100x OCP 1


OCP 2
B
20x + 100x = 110 – 40 40

120x = 70 A
x = 7/12 20 C
10

P(1)
0 1
Step 7. State your decision to minimize cost.

SON 2 SON 1
A: 0 ≤ P(1) < 0.25 120
B: 0.25 < P(1) < 7/12 110

C: 7/12 < P(1) ≤ 1.0


D: Never optimal 90 90

If probability of SON 1 is 60

less than 0.25, choose A; OCP 1


OCP 2
B
if between 0.25 and 7/12, 40

choose B; A
if greater than 7/12, 20 C
choose C. 10

P(1)
0 1

You might also like