MIE 353 Engineering Economics Todays Goals

1
MIE 353 Engineering EconomicsTodays Goals

• Calculate the value of information
• HW 7, Due Wednesday Nov. 15. problem 18.27,
  18.31, 18.34
• Project presentations Thursday Dec. 13
• Have presentation ready for practice by Monday
  Dec. 10 ( 2 weeks from today)
• Most groups have a lot of work to do.

2
Project Drafts

• Two teams need to re-do their drafts. You need
  more than a paragraph describing your excel
  sheet.
• Careful about how you report your results. Teams
  are reporting a present worth of x when really
  it is a present cost. Find a way to communicate
  this the total lifecycle cost in todays
  dollars is

3
Project

• Executive summary (not to exceed 1 page)
• Introduction
• Problem formulation study analysis
• Values objective state what is important to
  decision maker and how you have formulated the
  objective (i.e. the mathematical representation)
• Alternatives clearly describe the alternatives
  considered
• Information Describe the information and where
  you got it. If there is a lot of data put data
  itself in appendix
• Study Analysis
• Present equivalent worths of each alternative
  (NPV)
• Present sensitivity analysis. Include figures
• Conclusions and recommendation
• Appendices (backup calculations and supporting
  documentation)

4
Project

• Executive Summary
• This is a summary of the report for an executive
  who is very busy. She must decide whether she
  needs to read the whole report or delegate. She
  should know enough to ask interesting questions.
• It is NOT a teaser. Definitely put your main
  conclusions here.

5
Project

• Problem Description and Problem Formulation
• Use both words and tables or figures as
  appropriate.
• Outline your report before you write it. Make the
  order of the sections sensible. Try two different
  organizational structures.

6
Project

• Study Analysis
• Present all information in a way that it is easy
  to interpret.
• To represent sensitivity on a figure, graph the
  PW of each alternative against the parameter you
  are uncertain about.

7
Project

• The Project must have no mathematical errors!
• Assign one person to read the report to make sure
  its sensible.
• Assign at least one person to go through with a
  calculator and check ALL numbers. Make sure that
  there is enough information in the report (
  appendix) so that all numbers can be checked.

8
Project

• The Project must have no mathematical errors!
• The Project must be well written.
• It must not have casual language.
• cheap a lot sort of
• It must not have any typos or grammatical errors.
• It must be clear, understandable, and concise.

9
Project

• Please review how to write a paragraph
• Topic sentence
• Supporting sentence
• Concluding sentence
• A paragraph is not a random collection of
  sentences.
• The most important thing is to think clearly.
  This is greatly helped by outlining or some other
  similar structuring device.

10
Group assignment

• You have 100M to invest for 2 years. You first
  invest for one year in either stocks or bonds
  then you can choose to keep the same investment
  or change after the first year.
• The probably of growth, recession, or depression
  in the first year is 0.7, .03, and 0. If growth
  the first year, then the probabilities remain
  same. If recession the first year, then the
  probabilities for 2nd year are 0.2, .07, .01. The
  table gives the rates of returns in each
  situation for each investment.
• Draw a decision tree and find the optimal
  investment policy.

11
Value of Information

• We may want to invest in better information.
  Before we do so, we should investigate how much
  value information has.
• Expected Value of Perfect Information

12
Uncertainty Analysis

• You want to buy a car. Car A costs 20,000 and is
  guaranteed for 5 years. Car B costs 12,000, but
  might break down. If it breaks, it will cost you
  15,000 to replace it (including hassle value).
  What should you do?
• If someone told you that car B would not break
  down for sure, what would you do?
• Buy car B at a cost of 12,000
• If someone told you that car B would break down
  for sure, what would you do?
• Buy car A at a cost of 20,000

13
Value of Perfect Information

• Lets say that you know a clairvoyant. If you pay
  them, they will tell you with certainty whether
  your car will break down or not. You need to
  decide if it is worth it.
• You dont know what the clairvoyant will tell
  you, but you know the probability that the car
  will break down.
• The probability that the car will break down is
  .22. Therefore, the probability that the
  clairvoyant will give you the information that
  the car will break down, is .22

14
Value of Information

• Think of it this way. Cars such as the one you
  are contemplating buying break down 22 of the
  time.
• The clairvoyant will correctly inform you when
  the car is going to breakdown.
• So, they will inform you that the car is going to
  break down 22 of the time.

15
Value of Information

• What is the probability that the clairvoyant will
  inform you that the car will not break down?
• It is 1-.22 .78.

16
Value of Information

• If you are going to get perfect information, then
  with probability .22 you will buy car A at a cost
  of 20,000 and with probability .78 you will buy
  car B at a cost of 12,000
• Your expected cost with perfect information is
  .22(20,000) .78(12,000) 13,760
• Your expected cost without information is 15,300
• So the EVPI is 15,300 - 13,760 1,540

17
Car A
Act then learn. This tree represents the problem
where you must decide before learning
20,000
15,300
15,300
Car B
18
Car A
Act then learn. This tree represents the problem
where you must decide before learning
20,000
15,300
15,300
Car B
Car A
20,000
Learn, then act. This tree represents the problem
where you learn before deciding. You know that
you will learn, but you dont know what you will
learn.
P.22
27,000
13,760
Car B
Car A
20,000
1-P.78
12,000
Car B
19
Car A
Act then learn. This tree represents the problem
where you must decide before learning
20,000
15,300
15,300
Car B
Car A
20,000
20,000
Learn, then act. This tree represents the problem
where you learn before deciding. You know that
you will learn, but you dont know what you will
learn.
P.22
27,000
13,760
Car B
Car A
20,000
1-P.78
12,000
12,000
Car B
20
Car A
Act then learn. This tree represents the problem
where you must decide before learning
20,000
15,300
15,300
Car B
Car A
20,000
20,000
Learn, then act. This tree represents the problem
where you learn before deciding. You know that
you will learn, but you dont know what you will
learn.
P.22
27,000
13,760
Car B
Car A
20,000
1-P.78
12,000
12,000
Car B
EVPI 15300 13,760 1,540
21
EVPI general method

• Calculate the expected value of the decision
  problem without information.
• Calculate the expected value of each alternative.
• Choose the alternative with the highest expected
  value.
• Calculate the expected value of the decision
  problem with information.
• For each possible outcome, choose the alternative
  whose value is highest.
• Calculate the expected value over all the
  possible outcomes.
• EVPI value of problem with information value
  of the problem without information

22
Example

• JJ is facing a choice between two deals. Both
  deals depend on the price of gas. Deal 1 is worth
  10 if gas goes up, and 5 if gas goes down. Deal 2
  is worth 12 if gas goes up, and 2 if gas goes
  down.
• The probability that gas goes up is 40.
• What is the EVPI for this problem?

23
Group work Example

• JJ is facing a choice between two deals. Both
  deals depend on the price of gas. Deal 1 is worth
  10 if gas goes up, and 5 if gas goes down. Deal 2
  is worth 12 if gas goes up, and 2 if gas goes
  down.
• The probability that gas goes up is 40.
• What is the EVPI for this problem?
• Draw a decision tree for the Act-then-learn
  problem. Calculate the expected value of this
  problem.
• Draw a decision tree for the Learn-then-act
  problem. Calculate the expected value of this
  problem.
• The EVPI is the difference between the two.

24
Uncertainty Analysis

• You want to buy a car. Car A costs 20,000 and
  has operating costs of 55 cents per mile. Car B
  costs 18,000, and has operating costs of 62
  cents per mile. You will keep the car for 5 years
  and your MARR is 10. What should you do?
• You have determined that you will drive 5000
  miles with probability .3 12000 miles with
  probability .6, and 20,000 with probability .1.
• Draw a decision tree for the problem with and
  without perfect information. Determine the EVPI

25
The payoff trees for each alternative
20,000 2.08(5000) 30,424
P.3
P.6
45,019
A
P.1
61,699
18,000 2.35(5000) 29,751
P.3
B
P.6
46,204
P.1
65,005
26
EVPI

• EVPI Eoutcome with perfect information
  Eoutcome without information