Scribd
Upload
609 views

Decision Tree Analysis

The document discusses decision tree analysis, which is a technique used in quantitative risk analysis. It involves constructing a decision tree diagram that shows the sequence of decisions and their expected monetary values (EMVs). Examples are provided to illustrate how to calculate EMV at each decision node and choose the alternative with the highest total EMV. Initial costs may also be considered by subtracting them from the total EMV. Key terms like decision nodes, chance nodes, and end branches are defined. The document concludes with an example decision tree analyzing the probability and impacts of holding an office picnic.

Uploaded by

dzuleffendi
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
609 views

Decision Tree Analysis

The document discusses decision tree analysis, which is a technique used in quantitative risk analysis. It involves constructing a decision tree diagram that shows the sequence of decisions and their expected monetary values (EMVs). Examples are provided to illustrate how to calculate EMV at each decision node and choose the alternative with the highest total EMV. Initial costs may also be considered by subtracting them from the total EMV. Key terms like decision nodes, chance nodes, and end branches are defined. The document concludes with an example decision tree analyzing the probability and impacts of holding an office picnic.

Uploaded by

dzuleffendi
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 10

Decision Tree Analysis

In the previous chapter, we took a look at Expected Monetary Value or EMV Analysis.
The Decision Tree Analysis is another tool/technique that we use in Quantitative Risk
Analysis that directly uses this EMV Analysis. In this chapter, we are going to take a
detailed look at Decision Tree Analysis
Decision Tree Analysis
Decision Tree Analysis is used to make decisions based on the risks that could impact
us in the various possible scenarios we may encounter in future. It calculates the
Expected Future Value of an activity based on the current impact & probability of all
risks.
Decision Tree Analysis uses a Decision Tree Diagram. In the tree, we start at the
starting point and go through the tree and take a decision based on the EMV for the
Alternatives that are available for us. It shows a sequence of inter-related decisions and
their respective EMVs so that you can take a good and properly thought-out decision.
Decision Tree Analysis is used typically to take decisions dealing with Time or Cost.
Let us now take a look at some examples to understand Decision Tree Analysis:
Example 1:
Let us say, we are given the task of deciding between Vendor A and Vendor B. Vendor A
has a Success Probability of 55% and an Impact of $ 70,000 while there is no impact on
Failure. Similarly Vendor B has a 75% probability of Success and has an impact of $
55,000 and he too has no impact on Failure. Based on this information, how would you
choose the Vendor?

The simple Answer would be Use Decision Tree Analysis. So, based on this question,
if I were to create a Decision Tree, it would look like below:

So, here:
EMV for Vendor A: = 70000 * 55% = 38,500
EMV for Vendor B = 55000 * 75% = 41,250

Now, you know the EMV for each vendor. So, the wiser choice would be to choose
Vendor B because the Expected Monetary Value of choosing Vendor B is greater than
Vendor A.

Trivia:
If we had just considered the Impact, Vendor A would look like a better choice because
he has a higher impact. But, he has a lower probability. So, Vendor B, even though has
a lower impact, is selected because the combination of both probability and impact
makes him the better choice.
Example 2:
Example 1 was an all positive scenario where there is no Impact for failure. What would
you do in an all Negative Scenario?? Look at the Decision Tree Below:

In this example, in both cases, there is no impact if the outcome is a success. But, both
Vendors A & B have an impact on failure and have a probability of failure too. So in this
case, you may be wondering why I chose Vendor B instead of A, even though the EMV
for A is higher. Are you???
In example 1, we were looking at positive EMV (For Success or Profits) so, we chose
the vendor with higher profitable EMV. Whereas, in this case we are calculating
Negative EMV (For Failure or Losses). So, choosing the vendor who would cause lower
losses in case of a failure would be a better choice. Wouldnt it??

Example 3: In examples 1 & 2, we took a look at trees that either had a positive or
negative impact only. What must we do if we have both? Look at the tree below:

In this case, both Vendors A & B have Impact on both Success & Failure. So, the EMV
for each vendor is the sum of the Individual EMVs.
For Vendor A:
EMV for Failure = 1000 * 30% = 300
EMV for Success = 6000 * 70% = 4200
Total EMV for Vendor A = $ 4,500/For Vendor B:
EMV for Failure = -1200 * 40% = -480

EMV for Success = 7500 * 60% = 4500


Total EMV for Vendor B = $ 4,020/So, based on the total EMV, Vendor A is the better choice
Trivia:
Did you note that Vendor B has a negative monetary impact in case of failure??? Be
careful and note the sign If you did the calculation in a hurry and ignore the
symbol, your EMV for Vendor B wouldve been $ 4,980 suggesting that Vendor B is the
better choice. Whereas, the truth was that, because of the ve impact on failure, Vendor
A is the better choice.
Example 4:
In all of the examples above, our decision was based solely on the impact and
probability of the scenarios outcome. What must we do in cases where the decision
should also take into account the initial expenses incurred for the activity?
Lets say, your company has grown hugely in the past couple of years and your current
office does not have enough space to accommodate the new guys. So, now you have
two choices Either to construct/purchase a new office or expand the current premises
to accommodate the newer guys. In either case, there is cost involved in completion of
the office premises. Plus, there could be a high demand for the new space which results
in high profits or there could be a low demand resulting in lower profits. This scenario is
outlined in the Decision Tree below:

In Trees where an initial investment is present, we proceed just like the other scenarios
wherein we calculate the EMV for each alternative and sum it up. After that, we deduct
the Initial Expenses to arrive at the actual monetary value of the alternative.
Trivia:
In simpler terms, lets say, I invest 5,000 rupees today and earn 10,000 after 6 months,
my profit is 5,000 whereas, if I invest 10,000 rupees and earn 12,000 at the end of 6
months, my profit is only 2,000. Though the eventual money I get at the end of 6 months
is higher in the second case, it also means that I invest a larger amount up front thereby
reducing profits. So, option 1 where in invest 5000 and get 10000 at the end of 6
months is the better choice. Isnt it?
So, the EMV Calculation works out as follows:

For Build:
Total EMV = 522,000
If we include Initial Cost Net EMV = $ 292,000/For Expand:
Total EMV = 393,000
If we include Initial Cost Net EMV = $ 298,000/So, the decision here would be to expand the current office premises. Even though the
profits that we may earn if we move to a new office are higher, there is a higher impact if
people are not willing to move and a high initial cost. As a result, the EMV of expanding
the current office is more profitable and hence it is selected.

Some Important Decision Tree Related Terms:


For the exam, actually speaking, whatever we have covered so far is more than
sufficient. But, for the same of completeness, I want to cover one last topic related to
Decision Trees. There are 3 key terms that we will use while using Decision Trees in
real life. They are:
a. Decision Node The Point where an action or decision needs to be made Signified
by a solid black square
b. Chance Node The Point where events that cannot be controlled by the person who
is taking the decision happen Signified by a solid black dot
c. End of Branch - The end point with nothing connected on one end Signified by a
solid tilted Triangle.
The sample Decision Tree below can help you understand better as to what I am
trying to convey here:

I repeat, knowing these terms or these symbols are not required for the Exam. But, we
are not studying just to become certified. Our aim is to become better Risk Managers.
So, knowing these will only help you perform your duties better

Prev: Expected Monetary Value Analysis


Next: Modeling and Simulation
Posted by Anand VijayaKumar at 4:08 AM
Labels: decision making, decision tree, decision tree analysis, make or buy decision,
project risk analysis, quantitative analysis decision tree analysis

Example of Impact Analysis / Decision Tree Analysis


Illustrated above is a sample of a decision making tree. Let us assume that a office picnic is
being planned and is dependent on the weather. The condition for deciding on the picnic, or
the probability of having the picnic should value 0.65 / 1 or 65% for the picnic to be held. If
the value is less that 65%, a decision not to have the picnic is taken.
In this illustration, assume that from data obtained from external sources, there is a 30%
chance of good weather and 70% chance of bad weather.
In good weather, there is a 60% chance that the picnic will be held and a 40% chance that
the picnic will not be held. In bad weather there is a 20% chance the picnic will be held and
an 80% chance the picnic will not be held.
In order to come to a decision, the probability of the entire tree will have to be calculated,
and if the value is 0.65 or above, then a picnic will be decided upon. If criteria are not met,
there will be no picnic.

Probability of Having an Office Picnic


= probability of an office picnic in good weather + probability of an office picnic in bad
weather
= [0.30 x 0.60] + [0.70 x 0.20]
= 0.18 + 0.14
=0.32
= 32% chance of having a picnic.
Since the criteria required for having a picnic is a minimum of 65%, it can easily be assumed
that no picnic will be held.

Conclusion

It can easily be seen that the probability of an outcome can be calculated when using the
decision making tree diagram in project management. It is easy to follow and the impact is
obtained in numerical values which help in decision values rather than assumptions.
* Image Credit: Diagram created by Amanda Dcosta

You might also like