WOLLO UNIVERSITY
KOMBOLCHA INSTITUTE OF TECHNOLOGY
SCHOOL OF CIVIL ENGINNERING
DEPARTMENT OF CONSTRUCTION TECHNOLOGY AND
MANAGEMENT
CONSTRACTION PERFORMANCE
NAME ID
1. ELENI MULUGETA 0862/13
2. HSNA ADEM 1352/13
3. ABRHAM NIGUSU 0127/13
4. ANWAR TEFERA 4008/13
5. TAGESSE SALO 3766/13
Submitted to Mr. Teshome k
Decision Tree Risk Analysis
Introduction
Decision-making is a critical process in various fields, including business, finance,
healthcare, and project management. Whether selecting an investment, developing
a strategic plan, or diagnosing a medical condition, professionals must navigate
uncertainty and assess potential risks. Making well-informed decisions requires
analyzing multiple factors, including possible outcomes, associated probabilities,
and potential rewards or consequences.
One of the most effective tools for risk analysis and decision-making is the
decision tree. A decision tree is a structured, graphical representation that maps out
different choices, their possible outcomes, and the likelihood of each scenario
occurring. By visually organizing information in a logical manner, decision trees
provide a clear framework for evaluating complex situations. They help individuals
and organizations systematically compare options, weigh risks, and make rational
decisions based on available data.
Decision trees are widely used in various industries due to their simplicity,
transparency, and ability to handle uncertainty. Businesses use them to optimize
financial strategies, project managers rely on them to assess potential risks, and
healthcare professionals apply them to guide patient treatment plans. Regardless of
the field, decision trees enhance decision-making by breaking down complicated
problems into manageable steps, ultimately leading to more effective and strategic
choices.
Understanding Decision Trees
A decision tree consists of the following key components:
1. Decision Nodes (Squares) – Represent points where a decision must be
made.
2. Chance Nodes (Circles) – Indicate uncertain outcomes that may occur with
certain probabilities.
3. Branches – Connect nodes and represent different decisions or events.
4. Outcome Nodes (Triangles or End Points) – Show the final results of each
possible path in the tree.
A decision tree starts with a single decision node that branches out into different
options. Each option leads to either another decision node, a chance node, or an
outcome node. By following different paths, analysts can evaluate the best course
of action.
Steps in Decision Tree Risk Analysis
1. Define the Decision Problem
The first step is to clearly state the problem that requires a decision. For instance, a
company might need to decide whether to launch a new product, invest in a
project, or enter a new market.
2. Identify Possible Choices
List all potential options available. For example, if a company is considering
launching a new product, its choices could be:
Launch the product immediately
Delay the launch for further market research
Cancel the launch
3. Determine Risks and Uncertainties
Each choice comes with risks and uncertainties. For example, launching the
product immediately might lead to market competition, uncertain customer
demand, and potential production issues. Risks can include:
Financial risks: Losses from a failed product launch.
Operational risks: Production bottlenecks, supply chain issues.
Market risks: Competitor reactions, customer preferences.
Regulatory risks: Compliance with industry laws.
4. Assign Probabilities to Outcomes
For each uncertain event, probabilities are assigned based on historical data, expert
opinions, or statistical analysis. These probabilities help quantify risk. If past
product launches have a 70% success rate, this probability can be applied.
5. Estimate Potential Outcomes and Payoffs
Each decision path has associated financial and non-financial consequences. These
include:
Revenue and profit estimations
Costs of marketing, production, and distribution
Customer response and market share
Reputational and operational risks
These factors are quantified to create a more structured decision-making
framework.
6. Calculate Expected Values
The expected value (EV) of each decision path is calculated using the formula:
EV = (Probability of Outcome 1 × Payoff) + (Probability of Outcome 2 ×
Payoff) +...
The decision with the highest EV is generally the best choice.
7. Make the Final Decision
Once all expected values are calculated, the best decision is selected based on
quantitative and qualitative considerations.
Example of Decision Tree Risk Analysis
Suppose a company is deciding whether to launch a new software product. The
options and risks are:
1. Launch Immediately:
o 70% chance of success (500,000 profit)
o 30% chance of failure (200,000 loss)
2. Delay for Market Research:
o 60% chance of success (600,000 profit)
o 40% chance of failure (100,000 loss)
Using the EV formula:
EV (Launch Immediately) = (0.7 × 500,000) + (0.3 × -200,000) = 350,000 -
60,000 = 290,000
EV (Delay for Research) = (0.6 × 600,000) + (0.4 × -100,000) = 360,000 -
40,000 = 320,000
Since delaying for research has a higher expected value, it is the preferable choice.
Advanced Considerations in Decision Tree Analysis
Sensitivity Analysis
Sensitivity analysis involves testing how changes in probabilities and payoffs
affect the decision tree. By adjusting input values, businesses can determine the
robustness of their decisions. This is particularly useful in volatile industries where
conditions change rapidly, allowing companies to prepare for multiple scenarios
and minimize risks.
Monte Carlo Simulation
Monte Carlo simulation is a statistical method used to model uncertainty in
decision tree analysis. It generates thousands of possible outcomes based on
probability distributions, helping organizations assess risk more comprehensively.
This method enhances the accuracy of risk assessment by considering a wide range
of potential variations and their impacts, reducing the likelihood of unexpected
failures.
Bayesian Decision Trees
Bayesian decision trees incorporate prior knowledge and dynamically update
probabilities as new information becomes available. This makes decision-making
more adaptive and data-driven. Bayesian approaches are particularly useful in
industries like finance and healthcare, where ongoing data collection allows for
continuous improvement in decision-making strategies.
Real-World Case Studies
1. Amazon’s Market Expansion: Amazon uses decision trees to analyze
market risks before expanding into new regions, considering customer
demand, logistics, and regulatory risks.
2. Pharmaceutical Industry: Drug companies use decision trees to determine
whether to invest in new medications, weighing R&D costs against potential
market success.
3. Financial Risk Assessment: Banks apply decision trees to assess loan
approvals and credit risks, predicting customer defaults based on historical
data.
Advantages of Decision Tree Analysis
1. Clear Visualization – Helps break down complex decisions into a
structured format, making it easier to understand different scenarios.
2. Quantifies Risks – Assigns probabilities to different outcomes, making risk
assessment more precise and data-driven.
3. Data-Driven Decision Making – Uses numerical values to make objective
decisions, reducing biases and improving strategic choices.
4. Identifies the Best Strategy – Helps businesses choose the option with the
highest expected return, maximizing profitability and efficiency.
5. Flexible and Adaptable – Can be updated as new information becomes
available, ensuring dynamic decision-making in changing environments.
6. Scenario Analysis – Allows companies to model different scenarios to test
potential business strategies, reducing uncertainty and improving
preparedness.
7. Simplifies Complex Problems – Breaks down multi-step decision-making
processes into smaller, manageable parts, making analysis more structured
and systematic.
8. Supports Machine Learning Applications – Decision trees are used in
artificial intelligence and machine learning to classify data, predict
outcomes, and optimize processes.
9. Cost-Effective Analysis – Compared to more complex statistical models,
decision trees provide an effective and intuitive way to analyze risks without
requiring advanced mathematical skills.
Limitations of Decision Tree Analysis
1. Dependency on Accurate Data – Requires precise probability estimates,
which may not always be available. Inaccurate or insufficient data can lead
to flawed decision-making, as the entire analysis relies on the quality and
reliability of input information.
2. Complexity – Large decision trees can become difficult to manage and
interpret. As the number of decision nodes and chance events increases, the
tree can grow excessively, making it harder to understand and increasing the
likelihood of errors.
3. Assumption of Independence – Assumes that outcomes are independent,
which is not always the case in real-world scenarios. In many cases, external
factors such as market trends, economic shifts, or competitive actions may
influence multiple outcomes simultaneously.
4. Risk of Overfitting – Overly detailed trees may consider unnecessary
variables, making the analysis less practical. Overfitting occurs when a
decision tree is too finely tuned to historical data, reducing its ability to
predict future events accurately.
5. Computational Limitations – As trees grow in complexity, computational
resources may be needed to process vast amounts of data. Larger decision
trees require more processing power and time, which can make real-time
decision-making difficult, especially in dynamic business environments.
6. Sensitivity to Small Changes – A slight variation in input values,
probabilities, or costs can significantly alter decision tree outcomes, making
them less reliable in highly uncertain environments.
7. Lack of Consideration for External Factors – Decision trees often focus
on direct outcomes and fail to account for indirect consequences, such as
social, ethical, or long-term strategic implications.
Conclusion
Decision tree risk analysis is a powerful and structured approach to handling
complex decision-making scenarios. By breaking down decisions into smaller
components, assigning probabilities, and calculating expected values, businesses
and individuals can evaluate risks and make informed choices. While decision
trees have limitations, they remain one of the most widely used tools for risk
assessment and strategic planning. With advancements such as Monte Carlo
simulations and Bayesian decision trees, the methodology continues to evolve,
providing more sophisticated and accurate risk assessment techniques.
Organizations that integrate decision tree analysis into their strategic frameworks
benefit from improved decision-making, risk mitigation, and long-term success.
