D AT E | M O N T H | Y E A R BU275: Business Decision

Presented by
First Name Last Name Models
Instructor – Dr Jason Hurley
WINTER 2023
 Decision Analysis (part 2)
D AT E | M O N T H | Y E A R

Presented by
First Name Last Name
Decision Analysis

• The field of decision analysis provides a framework for making

important decisions
• Decision analysis helps us make a decision from a set of
possible decision alternatives when uncertainty/risk regarding
D AT E | M O N T H | Y E A R

the future
Presented byexists
First Name Last Name

The goal of decision analysis is to make the optimal decision

Decision Making Criteria

Five commonly used criteria:

1. Optimistic approach (Maximax or Minimin)
Uncertainty
2. Pessimistic (Conservative) approach (Maximin or Minimax)
(without probabilities)
3. Minimax Regret approach
D AT E | M O N T H | Y E A R
4. Equal Likelihood (Laplace) approach
Presented by
5. Expected Value
First Name Last approach
Name
Risk
(with probabilities)
… and a sixth decision criterion:
6. Expected Opportunity Loss
Additional Consideration

Dominated Alternatives
• If one decision alternative is always as good as or better than
another, it dominates the other

D AT E | M O N T H | Y E A R
• If one decision alternative is always better than another, it
Presented by
absolutely
First Name dominates
Last Name the other

• Drop dominated decision alternative from further consideration

Equal Likelihood (Laplace) Approach

The Equal Likelihood (Laplace) criterion assumes that each state of nature is

equally likely to occur
• Multiply the decision payoff for each state of nature by an equal weight
• Select the decision alternative with the greatest (least) weighted payoff value
for
D AT a
E |maximization
M O N T H | Y E A R (minimization) problem

Presented by
First Name Last Name State of Nature

Alternative Oil Dry

Drill for Oil 700,000 -100,000

Sell the Land 90,000 90,000

Probability 0.5 0.5

Equal Likelihood (Laplace) Approach

State of Nature

Alternative Oil Dry Weighted Payoff

D AT E | M O N T H | Y E A R
Drill for Oil 700,000 -100,000 300,000
Presented by
First Name Last Name
Sell the Land 90,000 90,000 90,000

Probability 0.5 0.5

Expected Value Approach
• The Expected Value (EV) approach can be used when probabilistic
information regarding the states of nature (outcomes) is available
• Procedure:
• Calculate the EV of each decision alternative by multiplying the
D AT E | M O N T H | Y E A R
respective payoff under each state of nature by the probability of that
Presented
state by
occurring, and then summing the products
First Name Last Name
• Select the decision alternative that yields the greatest (least) EV for a
maximization (minimization) problem

Note: EV is the average payoff that would result if the same decision situation occurred many
times (Law of Large Numbers), not the payoff that would result from a single decision.
Expected Value Approach

The expected value (EV) of a decision alternative (di) is defined as:

D AT E | M O N T H | Y E A R

where: Presented by
First Name Last Name
N = the number of states of nature
P(sj ) = the probability of state of nature sj
Vij = the payoff corresponding to decision alternative di and
state of nature sj
Expected Value Approach

State of Nature

Alternative Oil Dry EV

Drill for Oil 700,000 -100,000 100,000

D AT E | M O N T H | Y E A R

Presented
Sell by
the Land 90,000 90,000 90,000
First Name Last Name

Probability 0.25 0.75

𝐸𝑉 ( 𝐷𝑟𝑖𝑙𝑙 ) =700,000 ( 0.25 ) −100,000 ( 0.75 )=¿

𝐸𝑉 ( 𝑆𝑒𝑙𝑙 ) =90,000 ( 0.25 ) +90,000 ( 0.75 )=¿
Expected Value Approach
Pros Cons
• Accounts for all states of • Uncertainty assigning
nature and their values to probabilities
probabilities
• Largely subjective in
• Expected payoff
D AT E | M O ≈R average
NTH | YEA
nature
payoff
Presented by
• It ignores a decision
• On First Name Last Name
average, repeatedly maker’s risk tolerance (risk
applying EV approach to averse or risk seeking)
make decisions will lead
to larger payoffs in the
long run than any other
criterion.
Expected Opportunity Loss (EOL) Approach

EOL is the expected regret for each decision

• Always choose alternative that minimizes EOL
Regret Table
State of Nature

D ATAlternative
E | MONTH | YEAR Oil Dry ER
Presented by
First Name
Drill forLast
Oil Name 0 190,000 142,500

Sell the Land 610,000 0 152,500

Probability 0.25 0.75

𝐸𝑂𝐿 ( 𝐷𝑟𝑖𝑙𝑙 )=0 ( 0.25 )+190,000 ( 0.75 )=¿

𝐸𝑂𝐿 ( 𝑆𝑒𝑙𝑙 ) =610,000 ( 0.25 )+ 0 ( 0.75 ) =¿
Decision Trees

• A decision tree is a chronological representation of the decision problem

• Each decision tree has two types of nodes:
• round (probability or event) nodes correspond to the states of nature
• square
D AT E |(decision)
M O N T H | Y E A Rnodes correspond to the decision alternatives

• The branches leaving each round node represent the different states of
Presented by
First Name Last Name
nature while the branches leaving each square node represent the
different decision alternatives
• At the end of each limb of a tree are the payoffs attained from the series
of branches making up that limb
The Goferbroke Company Example

Step 1 – begin the tree with a decision (square) node and use
branches to define all decision alternatives

D AT E | M O N T H | Y E A R

Presented by
First Name Last Name
The Goferbroke Company Example

Step 2 – define all possible outcomes/states of nature for each

decision alternative using outcome (round) nodes and branches

D AT E | M O N T H | Y E A R

Presented by
First Name Last Name
 The Goferbroke Company Example
Step 3 – calculate the expected value for each combination of
decision alternative and state of nature (outcome [round] node)
Probabilities
EV(Drill) =
(0.25*700000) +
(0.75*-100000)
D AT E | M O N T H | Y E A R

Presented by
Expected Payoff =First Name Last Name
Max(100000,90000) Payoffs
CHOOSE TO
DRILL AS IT
YIELDS
HIGHEST EV EV(Sell) =
(0.25*90000) +
(0.75*-90000)
Expected Value of Perfect Information (EVPI)

• Expected value without perfect information

• The expected payoff of the best decision when you use the EV approach

• Expected value with perfect information

• Create a new row in the table which lists the best payoff for each state of nature
D AT E | M O N T H | Y E A R

• Determine thebyexpected payoff of this row to obtain expected payoff with perfect information
Presented
First Name Last Name
• Expected Value of Perfect Information (EVPI)
• EVPI = The difference between expected value with perfect information and expected
value without perfect information
• Note that EVPI = EOL for the best decision alternative (i.e., minimum EOL)

• Always positive
Obtaining More Information

• Is it possible to spend money to acquire more information (i.e.,

better estimates of the true state of nature that will occur)?

• If so, is the expenditure worthwhile?

D AT E | M O N T H | Y E A R

Presented by
• Check
Firstby calculating
Name Last Name EVPI and comparing to the cost of obtaining the
information
• If EVPI > Cost, then it may be worthwhile to obtain more info
• If EVPI < Cost, then it is not worthwhile to obtain more info
 Decision Tree with Perfect Information
• Obtain “perfect information” about the state of nature before making a
decision – reverse the chronological order of outcome and alternatives

D AT E | M O N T H | Y E A R

Presented
Expected value WITH perfectby Expected Payoff =
information (EVPI)First
= Name Last Name Max(700,000,90,000)
(0.25*700,000)+(0.75*90,000) = 

Expected Payoff =
Max(-100,000,90,000)
 Expected Value of Perfect Information (EVPI)

State of Nature

Alternative Oil Dry EV

Drill for Oil 700,000 -100,000 100,000

D AT E | M O N T H | Y E A R
Sell the Land 90,000 90,000 90,000
Presented by
EV
Firstwith
NamePI Last Name 700,000 90,000 242,500

Probability 0.25 0.75

𝑬𝒙𝒑𝒆𝒄𝒕𝒆𝒅𝑽𝒂𝒍𝒖𝒆𝑶𝑭 𝑷𝒆𝒓𝒇𝒆𝒄𝒕 𝑰𝒏𝒇𝒐𝒓𝒎𝒂𝒕𝒊𝒐𝒏(𝑬𝑽𝑷𝐼)=𝐸𝑉 𝑤𝑖𝑡h 𝑃𝐼−𝐸𝑉 𝑤𝑖𝑡h𝑜𝑢𝑡𝑃𝐼=242,500−100,000=𝟏𝟒𝟐,𝟓𝟎𝟎

𝑵𝒐𝒕𝒆 𝒕𝒉𝒂𝒕 𝑬𝑽𝑷𝑰 =𝟏𝟒𝟐 , 𝟓𝟎𝟎= 𝑬𝑶𝑳( 𝑫𝒓𝒊𝒍𝒍)
Relationship Between EVPI and EOL

EV without perfect information + EOL for each decision alternative = EV

with perfect information

Recall that EV (Drill) = 100,000 and EV (Sell) = 90,000 and regret table:
D AT E | M O N T H | Y E A R

Presented by State of Nature

First Name Last Name EV + EOL = 242,500
Alternative Oil Dry EOL

Drill for Oil 0 190,000 142,500 Drill: 100,000 + 142,500 =

Sell the Land 610,000 0 152,500
242,500

Probability 0.25 0.75 Sell: 90,000 + 152,500 =

242,500
 Thanks for your attention!

Next class:
• Decision Analysis continued…
D AT E | M O N T H | Y E A R

• Bayesian
Presented Updating
by
• First Name
Decision Last Name
making with sample information
• Expected value of sample information
• Efficiency of sample information