0% found this document useful (0 votes)
327 views10 pages

Exam Questions: Decision Tree Problems

This document contains a summary of several decision problems involving decision trees. For the first problem, the optimal decision is to order 100 books using the Laplace criterion, with an expected monetary value of $740. For the second problem on yogurt sales, the optimal decision is to buy 300 cups, with an expected monetary value of $62.5. The third problem involves a payoff table for decisions A, B and C, where the optimal decision is B with an expected value of $511.

Uploaded by

DiamonetteSyncon
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
327 views10 pages

Exam Questions: Decision Tree Problems

This document contains a summary of several decision problems involving decision trees. For the first problem, the optimal decision is to order 100 books using the Laplace criterion, with an expected monetary value of $740. For the second problem on yogurt sales, the optimal decision is to buy 300 cups, with an expected monetary value of $62.5. The third problem involves a payoff table for decisions A, B and C, where the optimal decision is B with an expected value of $511.

Uploaded by

DiamonetteSyncon
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 10

DECISION TREE PROBLEMS

Exam Questions
Question I
Waldo Books

States of Nature
Number of Books Demanded
Amount Ordered 50 100 150 200
50 $500 $500 $500 $500
100 ($300) $1,000 $1,000 $1,000
150 ($1,100) $200 $1,500 $1,500
200 ($1,900) ($600) $700 $2,000

Amt Ordered (b) Maximax (c) Maximin (d) LaPlace Decisions:


50 $500 $500 $500 (b) Order 200 books for Maximax
100 $1,000 ($300) $675 (c) Order 50 books for Maximin
150 $1,500 ($1,100) $525 (d) Order 100 books for LaPlace
200 $2,000 ($1,900) $50 (e) Order 200 books for Hurwicz

(e) Hurwicz: α = 0.7; 1-α = 0.3


Order 50: 0.7(500) + 0.3(500) = 500
Order 100: 0.7(1000) + 0.3(-300) = 610
Order 150: 0.7(1500) + 0.3(-1100) = 720
Order 200: 0.7(2000) + 0.3(-1900) = 830
Regrets
States of Nature
Amount
Ordered 50 100 150 200 Maximum
50 $0 $500 $1,000 $1,500 $1,500
100 $800 $0 $500 $1,000 $1,000
150 $1,600 $800 $0 $500 $1,600
200 $2,400 $1,600 $800 $0 $2,400
Decision: Order 100 for Minmax regret
Expected values:
Order 50: 0.2(500)+0.35(500)+0.25(500)+0.2(500) = 500
Order 100: 0.2(-300)+0.35(1000)+0.25(1000)+0.2(1000) = 740
Order 150: 0.2(-1100)+0.35(200)+0.25(1500)+0.2(1500) = 525
Order 200: 0.2(-1900)+0.35(-6200)+0.25(700)+0.2(2000) = -15
Decision: Order 100; Expected Monetary Value = 740

EVPI = Either Min expected Regret or EWPI – EWOPI; Min Expected Regret = 485
Page 1 of 10
b) EWPI = 0.2(500)+0.35(1000)+0.25(1500)+0.2(2000) = 1225; EWOPI = 740; EVPI = 1225– 740 = 485

Question II: Yogurt Hut

Decision Maximax Maximin Hurwicz LaPlace


Buy 200 50 50 50 50
Buy 300 75 25 55 58.33
Buy 400 100 0 60 50

Hurwicz: Buy 200: 0.6*50 + 0.4*50 = 50


Buy 300: 0.6*75 + 0.4*25 = 55
Buy 400: 0.6*100 + 0.4*0 = 60
Decisions: Maximax = Buy 400; Maximin = Buy 200; Hurwicz = Buy 400; LaPlace = Buy 300
e) Minmax Regret
Regrets
Decision Sell 200 Sell 300 Sell 400 Max
Buy 200 0 25 50 50
Buy 300 25 0 25 25
Buy 400 50 25 0 50
Decision: Buy 300

f) Expected Monetary Values: Buy 200: 0.25*50 + 0.45*50 + 0.30*50 = 50.0


Buy 300: 0.25*25 + 0.45*75 + 0.30*75 = 62.5 = EWOPI
Buy 400: 0.25*0 + 0.45*50 + 0.30*100 = 52.5

Decision: Buy 300; EMV = 62.5


g) Minimum EOL: Buy 200: 0.25*0 + 0.45*25 + 0.30*50 = 26.25
Buy 300: 0.25*25 + 0.45*0 + 0.30*25 = 13.75 EVPI = 13.75
Buy 400: 0.25*50 + 0.45*25 + 0.30*0 = 23.75

h) EVPI = EWPI – EWOPI


EWPI = 0.25*50 + 0.45*75 + 0.30*100 = 76.25; EVPI = 76.25 – 62.5 = 13.75

*****=====+++++

Question III

A decision maker has determined a payoff table for three decision alternatives A, B, and C as well as three
states of nature X, Y and Z:

Page 2 of 10
Decision Maximax Maximin LaPlace Hurwicz
A 750 -100 350 495
B 700 200 490 550
C 600 400 500 540

(a) For Maximax, the decision is A


(b) For Maximin, the decision is C
(c) For LaPlace, the decision is C

(d) Hurwicz: α = 0.7; 1-α = 0.3


A: 0.7(750) + 0.3(–100) = 495
B: 0.7(700) + 0.3(200) = 550
C: 0.7(600) + 0.3(400) = 540 The decision is B

(e) Regrets
Decision X Y Z Maximum
A 0 170 500 500
B 50 0 200 200
C 150 70 0 150
The decision is C: (4 marks)

(f) A: 0.4(750) + 0.3(400) + 0.3(–100) = 390


B: 0.4(700) + 0.3(570) + 0.3(200) = 511
C: 0.4(600) + 0.3(500) + 0.3(400) = 510

The optimal expected value is 511 and the decision is B

(g) A: 0.4(0) + 0.3(170) + 0.3(500) = 201


B: 0.4(50) + 0.3(0) + 0.3(200) = 80
C: 0.4(150) + 0.3(70) + 0.3(0) = 81

The EVPI is 80
(i)
EVPI = EWPI – EWOPI
EWPI = 0.4(750) + 0.3(570) + 0.3(400) = 591
EWOPI = 511 (Maximum EMV)
EVPI = 591 – 511 = 80

Decision Trees

Question IV: A Canadian company called Rainville


Required:
P(O) = P(O|D)P(D) + P(O|S)P(S) + P(O|G)P(G) = 0.8*0.4 + 0.3*0.4 + 0.1*0.2 = 0.46. So P(C) = 0.54
P(D|O) = P(O|D)P(D)/P(O) = 0.70; P(S|O) = P(O|S)P(S)/P(O) = 0.26; P(G|O) = P(O|G)P(G)/P(O) = 0.04
P(D|C) = P(C|D)P(D)/P(C) = 0.15; P(S|C) = P(C|S)P(S)/P(C) = 0.52; P(G|C) = P(C|G)P(G)/P(C) = 0.33

See Tree Below

Page 3 of 10
Folding Back

EV(6) = 0.4*–400 + 0.4*350 + 0.2*1,600 = 300


EV(7) = 0.7*–430 + 0.26*320 + 0.04*1,570 = –155
EV(8) = 0.15*–430 + 0.52*320 + 0.33*1,570 = 620

At node 3 the decision is “Drill”; fold back 300


At node 4 the decision is “Do not drill”; fold back –30
At node 5 the decision is “Drill”; fold back 620

EV(2) = 0.46*–30 + 0.54*620 = 321

At node 1 the decision is “Seismic Tests”; fold back 321

Optimal EMV = 321 million.


Decision: Do Seismic tests, if Open structure, do not drill, if closed structure, drill.

(c)
EVSI = EWSI – EWOSI ;
EVSI = 321 + 30 – 300 = 51

(d)
EVPI = EWPI – EWOPI
EWPI = 0.4*0 + 0.4*350 + 0.2*1600 = 460;
EWOPI = 300
EVPI = 460 – 300 = 160

(e)
Efficiency = EVSI/EVPI = 51/160 = 31.875%

Page 4 of 10
Question V
Page 5 of 10
A company is planning a plant expansion.

The drawing of the tree will show that we need the following probabilities: P(F); P(U); P(H|F); P(L|F); P(H|U) and
P(L|U).
 P(F) = P(F|H)*P(H) + P(F|L)*P(L) = 0.84*0.69 + 0.16*0.31 = 0.63; P(U) = 1 – P(F) = 0.37
P ( F | H ) P( H ) 0.84*0.69
 P( H | F ) = = = 0.92 ; thus P(L|F) = 0.08
P( F ) 0.63
P(U | H ) P( H ) 0.16*0.69
 P( H | U ) = = = 0.30 thus P(L|U) = 0.70
P(U ) 0.37

Page 6 of 10
EVSI = EWSI – EWOSI = 83.72 – 63.2 = 20.52 (million)
EVPI = EWPI – EWOPI = 150*0.69 + -$30*0.31 – 63.2 = 94.20 – 63.2 = 31 (million)
Efficiency = 20.52/31 *100 = 66.19%

Question VI
Bakery Products is considering the introduction of a new line of products.
P(G) = P(G|F)*P(F) + P(G|U)*P(U) = 0.8*0.5 + 0.4*0.5 = 0.6; therefore P(B) = 0.4
Page 7 of 10
P (G | F ) P ( F ) 0.4
P( F | G ) = = = 0.67; P (U | G ) = 1 - P ( F | G ) = 0.33
P (G | F ) P( F ) + P(G | U ) P(U ) 0.6
P( B | F ) P( F ) 0.2*0.5
P( F | B) = = = 0.25; P (U | B ) = 1 - 0.25 = 0.75
P ( B | F ) P( F ) + P( B | U ) P (U ) (0.2)(0.5) + (0.6)(0.5)

Page 8 of 10
EVSI = 17,380 + 5,000 – 10,000 = 12,380
EVPI = EWPI – EWoPI; EWPI = 0.5*100,000 + 0.5*0 = 50,000; EWOPI = 10,000; EVPI = 40,000
Efficiency = 12,380/40,000 = 30.95%

Question VII: A contractor is invited to bid on a construction job.


P(F) = P(F|G)* P(G) + P(F|B)* P(B) = 0.7*0.2 + 0.2*0.8 = 0.3. P(U) = 1 – P(F) = 0.7
P(G|F) = P(F|G)* P(G)/P(F) = 0.14/0.3 = 0.47; P(B|F) = 0.53;
P(G|U) = P(U|G)* P(G)/P(U) = 0.06/0.7 = 0.09 so P(B|U) = 0.91

Page 9 of 10
Decision: Do not buy the forecast and Bid. Expected Profit = $20,000

EVSI = 0
EVPI = EWPI – EWOPI
EWPI = 0.2*500,000 + 0.8*0 = 100,000
EWOPI = 20,000
EVPI = 80,000
Efficiency = 0

Page 10 of 10

You might also like