Probability & Statistical Modelling

(AQ077-3-2-PSMOD Version 1)

Decision Making Techniques

 Topic & Structure of the lesson

 Introduction
 Decision making Under Uncertainty
 Expected Value / Expected Opportunity
Loss
 Expected Value of perfect information
 Decision Trees

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <2> of 71
 Learning Outcomes

 At the end of this topic, You should be able to:

 Understand how to calculate and use Expected

Value
 Know how to value Perfect and Imperfect
information
 Use various decision rules
 Know what is meant by a Decision Tree
 Be able to describe Decision Trees and Outcomes
Nodes

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <3> of 71
  Understand how to draw Decision Trees using the
Forward Pass
 Know that the outcome values are calculated
using the backward pass
 Be able to incorporate Bayes’ Theorem into a
Decision Tree.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <4> of 71
 Key Terms you must be able to use

If you have mastered this topic, you should be able to use the following terms
correctly in your assignments and exams:
(Prepare your own list)

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <5> of 71
 Introduction

Decision making involves the following

steps:
 Recognise and clearly define the problem
 Collect the information needed to analyse
possible alternatives
 Choose and implement the most feasible
alternative

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <6> of 71
 Problem formulation
 A decision problem is characterized by decision
alternatives, states of nature, and resulting payoffs.
 The decision alternatives are the different possible
strategies the decision maker can employ.
 The states of nature refer to future events, not under
the control of the decision maker, which may occur.
States of nature should be defined so that they are
mutually exclusive and collectively exhaustive.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <7> of 71
 Decision Theory
a general approach to decision making
when the outcomes associated with
alternatives are often in doubt
 List the feasible alternatives
 List the events
 Calculate the payoff
 Estimate the likelihood of each event
 Select a decision rule

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <8> of 71
  When a decision has to be made, there
will be a range of possible action
 Each action will have a certain
consequences or payoff

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <9> of 71
 Decision Making Under Certainty
Manager knows which event will occur
pick the alternative with the best payoff
Possible Future Demand
Alternative Low High
Small facility 200 270
Large facility 160 800
Do nothing 0 0

What is the best choice if future demand will be low?

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <10> of 71
 Decision Making Under Uncertainty

Decision making without probability

(no probability of occurrence are assigned)

Decision making with probabilities

(probabilities can be assigned)

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <11> of 71
 Decision Making Without Probabilities

Decision Criteria
 Maximax criterion
 Maximin criterion
 Minimax regret criterion
 Hurwicz criterion

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <12> of 71
 Maximax Criterion
 maximise the maximum profit
 optimistic - looks at the best possible
payoffs
 ignores the possibility of a potential loss

N/B: maximax criterion deals with profit, if the payoff

table consists of costs, then minimin criterion would
be used ( i.e. minimise the minimum costs)

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <13> of 71
 Quick Review Question
Decision making without probability
(no probability of occurrence are assigned)
State of Nature

Decision Good Poor Economic

(purchase) Economic conditions
conditions
Apartment building $50,000 $30,000

Office building 100,000 -40,000

Warehouse 30,000 10,000

Determine the maximax criterion for our investment.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <14> of 71
 Quick Review Question
Decision making without probability
(no probability of occurrence are assigned)

Decision
(purchase)
Apartment building $50,000

Office building 100,000

Warehouse 30,000

The maximax criterion for our investment is to purchase office

building.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <15> of 71
 Maximin Criterion
looks at the worst possible payoffs
pessimistic and conservative
results in the maximum of the minimum
profit

Its equivalent is minimax (i.e. minimise the maximum

cost)

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <16> of 71
 Quick Review Question
Decision making without probability
(no probability of occurrence are assigned)
State of Nature

Decision Good Poor Economic

(purchase) Economic conditions
conditions
Apartment building $50,000 $30,000

Office building 100,000 -40,000

Warehouse 30,000 10,000

Determine the maximin criterion for our investment.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <17> of 71
 Quick Review Question
Decision making without probability
(no probability of occurrence are assigned)

Decision
(purchase)
Apartment building 30,000

Office building -40,000

Warehouse 10,000

The maximin criterion for our investment is to purchase apartment

building.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <18> of 71
 Minimax Regret Criterion
 Regret (opportunity loss) is the difference
between the payoff from the best decision
and all other decision payoffs in those
circumstances

 In this criterion decision maker attempts to avoid

regret by selecting the decision alternative that
minimises the maximum regret

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <19> of 71
 Quick Review Question
Decision making without probability
(no probability of occurrence are assigned)
State of Nature

Decision Good Poor Economic

(purchase) Economic conditions
conditions
Apartment building $50,000 $30,000

Office building 100,000 -40,000

Warehouse 30,000 10,000

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <20> of 71
 Quick Review Question
Decision making without probability
(no probability of occurrence are assigned)

Regret Table
State of Nature

Decision Good Poor Economic

(purchase) Economic conditions
conditions
Apartment building $50,000 0

Office building 0 70,000

Warehouse 70,000 20,000

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <21> of 71
 Quick Review Question
Decision making without probability
(no probability of occurrence are assigned)

Decision:
Decision (purchase)

Apartment building $50,000

Office building 70,000

Warehouse 70,000

According to minimax regret criterion, the decision should be

to purchase apartment building because this decision will
result, at most, $50,000 in regret.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <22> of 71
 Hurwicz Criterion
 a comprise between the maximax and
maximin criteria
 introduce , the coefficient of optimism
 0<  <1; It is a measure of the decision
maker’s optimism
 Multiplies the best payoff by  and the
worst by (1 - ) , then select the decision
which gives the maximum payoff

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <23> of 71
 When  = 0 it is effectively the maximin
criterion
when  = 1, it is the maximax criterion
 = 0.5 means the states of nature are
equally likely to occur, it is known as
Equally Likelihood Criterion

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <24> of 71
 Quick Review Question
State of Nature

Decision Good Poor Economic

(purchase) Economic conditions
conditions
Apartment building $50,000 $30,000

Office building 100,000 -40,000

Warehouse 30,000 10,000

Determine the Hurwicz criterion decision of our

investment example, let  = 0.4

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <25> of 71
 Quick Review Question

 = 0.4, (1- ) =0.6

Decision (purchase) Values

Apartment building

Office building

Warehouse

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <26> of 71
 Quick Review Question

Decision Values
(purchase)
Apartment building $50,000(0.4) + $30,000(0.6) = 38,000

Office building 100,000(0.4)+(-40,000)(0.6)= 16,000

Warehouse 30,000(0.4)+10,000(0.6)= 18,000

According to Hurwicz criterion ( =0.4), the

maximum weighted value is $38,000. Thus, the
decision would be to purchase the apartment
building.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <27> of 71
 Quick Review Question
State of Nature

Decision Good Poor Economic

(purchase) Economic conditions
conditions
Apartment building $50,000 $30,000

Office building 100,000 -40,000

Warehouse 30,000 10,000

Determine the best decision using the equal

likelihood criterion in our investment.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <28> of 71
 Quick Review Question

Decision Values
(purchase)
Apartment building

Office building

Warehouse

Determine the best decision using the equal

likelihood criterion in our investment.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <29> of 71
 Quick Review Question

Decision Values
(purchase)
Apartment building $50,000(0.5) + $30,000(0.5) = 40,000

Office building 100,000(0.5)+(-40,000)(0.5)= 30,000

Warehouse 30,000(0.5)+10,000(0.5)= 20,000

In applying the equal likelihood criterion, we are assuming

50% chance, that either state of nature will occur. Since
$40,000 is the highest weighted value, the decision would
be to purchase the apartment building.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <30> of 71
 Summary of criteria results:

Criterion Decision (purchase)

Maximax Office building

Maximin Apartment building

Minimax regret Apartment building

Hurwicz Apartment building

Equal likelihood Apartment building

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <31> of 71
 Conclusion
 Which criterion is appropriate is dependent
on the risk personality and philosophy of
the decision maker.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <32> of 71
 Example 1
Tutorial 2 Q2
A farmer in Georgia must decide which crop to plant next year
on his land: corn, peanuts, or soyabeans. The return from each crop will be
determined by whether a new trade bill with Russia passes the Senate. The
profit the farmer will realize from each crop given the two possible results on
the trade bill as shown in the following payoff table. Determine the best crop
to plant using the following decision criteria.
Trade Bill
Crop Pass ($) Fail ($)
Corn 35000 $8000
Peanuts 18000 12000
Soyabeans 22000 20000
(a) Maximax (d) Hurwicz ( α = 0.3)
(b) Maximin (e) Equal likelihood
(c) Minimax regret

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <33> of 71
 Decision Making with Probabilities

Expected Value Approach

 If probabilistic information regarding the states of
nature is available, one may use the expected value
(EV) approach.
 Expected value is computed by multiplying each
decision outcome under each state of nature by the
probability of its occurrence
 The decision yielding the best expected return is
chosen.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <34> of 71
 Payoff Tables
 The consequence resulting from a specific combination
of a decision alternative and a state of nature is a
payoff.
 A table showing payoffs for all combinations of decision
alternatives and states of nature is a payoff table.
 Payoffs can be expressed in terms of profit, cost, time,
distance or any other appropriate measure.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <35> of 71
 Quick Review Question

Payoff Table for the real estate investments

Calculate the expected value of each choice and

make the best decision.

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <36> of 71
 Expected Opportunity Loss
 Is the expected value of the regret for each
decision.
 Regret is the difference between the payoff from the
best decision and all other decision payoffs.
 Computed by multiplying each decision outcome
(probability) by the regret i.e. opportunity loss

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <37> of 71
 Quick Review Question

Payoff Table for the real estate investments

Calculate the expected opportunity loss (EOL) that

would be experienced by the decision maker ?

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <38> of 71
  An investor must decide among an apartment
building, an office building and a warehouse using
expected opportunity loss.
The regret values for each decision outcome were
shown below with the probabilities of occurrence
for each state of nature.

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <39> of 71
 The expected opportunity loss for each decision is
computed as follows:
EOL (apartment) = $50000(0.6) + 0(0.4) = $30000
EOL (Office) = $0(0.6) + 70000(0.4) = $28000
EOL (Warehouse) = $70000(0.6) + 20000(0.4) = $50000
The best decision results from minimizing the
expected regret or opportunity loss. Since $28000 is
the minimum expected regret, the decision is to
purchase the office building (This decision will result,
at most, $28,000 in regret)

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <40> of 71
 Expected value of Perfect information(EVPI)
It is the maximum amount a decision
maker would pay for addition information
Equals to the expected value given
perfect information minus the expected
value without perfect information.

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <41> of 71
 Quick Review Question

Payoff table with Decisions, given perfect information

is as below:

Determine the expected value of perfect

information.

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <42> of 71
 Expected value given perfect information
= $100,000 (0.6) + $30,000(0.4) = $72,000
Expected value without perfect information
= $100000 (0.6) – $40000(0.4) =$44000
EVPI = $72000 - $44000 = $28000 (lowest EOL)
(This is the maximum amount that the investor would
pay to purchase perfect information from other
source, such as an economic forecaster)

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <43> of 71
 Example 2

• The Loebuck Grocery must decide how many

cases of milk to stock each week in order to meet
demand. The probability distribution of demand
during a week is shown in the table below:

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <44> of 71
 Each case costs the grocer $10 and sells for $12.
Unsold cases are sold to a local farmer (who mixes the
milk with feed for livestock) for $2 per case. If there is a
shortage, the grocer considers the cost of customer ill
will and lost profit to be $4 per case. The grocer must
decide how many cases of milk to order each week.
(a) Construct the payoff table.
(b) Compute the expected value of each
alternative amount of milk that could be
stocked, and select the best decision.

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <45> of 71
 Decision Trees

 A decision tree is a chronological

representation of the decision problem. It is
a graphical device that forces the decision-
maker to examine all possible outcomes,
including unfavorable ones.
 makes easier the computation of the
expected values
 easy to understand the process of making
decision

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <46> of 71
  Each decision tree has two types of
nodes; round nodes correspond to the
states of nature while square nodes
correspond to the decision alternatives.
 The branches leaving each round node
represent the different states of nature
while the branches leaving each square
node represent the different decision
alternatives.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <47> of 71
  At the end of each limb of a tree are the
payoffs attained from the series of
branches making up that limb.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <48> of 71
 How to draw a decision Tree ?

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <49> of 71
AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <50> of 71
  Start a decision tree with a decision that
needs to be made which is represented by a
small square.
 From this square draw out lines towards the
right for each possible solution, and write
that solution along the line.
 Keep this lines as far apart as possible so
that you can expand on your thoughts.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <51> of 71
  At the end of each solution line, consider the
results:
 If the result of taking that decision is
uncertain, draw a circle.
 If the result is another decision that needs to
be made, draw another square.
 Write the decision or factor to be
considered above the square or circle.
 If you have completed the solution at the
end of the line, just leave it blank.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <52> of 71
  Starting from the new decision squares on
your diagram, draw out lines representing
the options that could be taken.
 From the circles draw out lines representing
possible outcomes.
 Again mark a brief note on the line saying
what it means.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <53> of 71
  Keep on doing this until you have drawn
down as many of the possible outcomes
and decisions as you can see leading on
from your original decision.
 Review your tree diagram. Challenge each
square and circle to see if there are any
solutions or outcomes you have not
considered. If there are, draw them in.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <54> of 71
  If necessary, redraft your tree if parts of it
are too congested or untidy.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <55> of 71
 Evaluate Your decision Tree
 Start by assigning a cash or numeric value to
each possible outcome – how much you
think it is worth.
 Next look at each circle and estimate the
probability of each outcome. (The total
should be 100% or 1)

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <56> of 71
 Calculating Tree Value
 Start on the right hand side of the decision
tree, and work back towards the left.
 Record the result after completing the
calculation on a node.
 All the calculations that lead to the result can
be ignored – effectively that branch of the
tree can be discarded. (Pruning the tree)

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <57> of 71
 Calculating the value of uncertain
outcome nodes
 Multiply the value of the outcomes by their
probability, and noting the result.
 The total value of that node of the tree is
gained by adding the results together.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <58> of 71
 Calculating the value of decision nodes
 Calculate the benefit of each decision (subtract
the cost from the value of that outcome that
you have already calculated)
 Select the decision which has the largest
benefit, and take that as the decision made

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <59> of 71
 Quick Review Question
Example:
 Burger Prince Restaurant is contemplating opening
a new restaurant on Main Street. It has three
different models, each with a different seating
capacity. Burger Prince estimates that the
average number of customers per hour will be 80,
100, or 120. The payoff table for the three models is
as follows:

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <60> of 71
  Calculate the expected value for each
decision. The decision tree on the next slide
can assist in this calculation. Here d1, d2, d3
represent the decision alternatives of models
A, B, C, and s1, s2, s3 represent the states of
nature of 80, 100, and 120.

CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <61> of 71
CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <62> of 71
CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <63> of 71
CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <64> of 71
CT042-3.5-2 Probability and Statistical Modeling Decision Making Techniques Slide <65> of 71
 Example 3

The financial success of the Downhill Ski Resort in the Blue Ridge Mountains
is dependent on the amount of snowfall during the winter months. If the
snowfall averages more than 40 inches, the resort will be successful; if the
snowfall is between 20 and 40 inches, the resort will receive a moderate
financial return; and if snowfall averages less than 20 inches, the resort will
suffer a financial loss. The financial return and probability given each level of
snowfall follows.

Snowfall Level (in) Prob Financial Return ($)

>40, 0.4 120,000
20-40, 0.2 40,000
<20, 0.4 - 40,000
A large hotel chain has offered to lease the resort for the winter for $40,000.
Draw a decision tree and use it to determine if the resort should operate or
lease. Explain your answer.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <66> of 71
 Follow Up Assignment

TUTORIAL QUESTIONS

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <67> of 71
 Summary of Main Teaching Points

 Elements common to decision theory

problems:
 An objective the decision maker is trying to
reach.
 Several courses of action
 A calculable measure of the benefit or worth
of the various alternatives
 Events beyond the control of the decision
maker
 Uncertainty concerning which outcome or
state of nature will actually happen.
AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <68> of 71
 Decision trees provide an effective method of
decision making because they:
 Clearly lay out the problem so that all choices can
be viewed, discussed and challenged
 Provides a framework to quantify of the values of
outcomes and the probabilities of achieving them
 Help us to make the best decisions on the basis of
our existing information and best guesses.

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <69> of 71
 Question and Answer Session

Q&A

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <70> of 71
 Next Session

Summary Measures of Statistics

AQ077-3-2 Probability and Statistical Modeling Decision Making Techniques Slide <71> of 71