Chapter 1
Uploaded by
silec99364Chapter 1
Uploaded by
silec99364Decision Science
From wartime logistics to modern boardrooms - a framework for making smarter, data-driven decisions under any condition.
Introduction to Decision Science
Understanding the origins, purpose, and building blocks of a discipline that replaces gut feeling with mathematical rigor.
Origin & Development
World War II Modern Mandate
Decision Science emerged from Operations Research to Now essential for ensuring financial stability and
solve complex logistics and resource allocation improving competitive advantage.
problems.
1 2 3
Scientific Method
Applies mathematical logic to management, replacing
intuition with structured analysis.
Types of Decisions & Models
Model Building
Structured Unstructured
Models are simplified representations of reality. You must
Routine, repeatable Complex, novel learn to identify decision variables and construct objective
decisions with clear decisions requiring functions that reflect organizational goals.
procedures. judgment and creativity.
Decision Making Under Uncertainty
When managers face the unknown without knowing exact probabilities, they rely on specific decision rules - each reflecting a
different management mindset.
Four Rules for the Unknown
Maximax Maximin
Optimistic: Choose the alternative with the highest Pessimistic: Choose the best of the worst-case
possible gain. scenarios.
Laplace Minimax Regret
Neutral: Treat all potential outcomes as equally likely. Regret-Averse: Minimize the opportunity loss from not
choosing the best option.
Comparing the Rules
Rule Mindset Strategy
Maximax Optimistic Maximize the maximum possible payoff
Maximin Pessimistic Maximize the minimum possible payoff
Laplace Neutral Average all outcomes equally
Minimax Regret Regret-Averse Minimize worst-case opportunity loss
Each rule suits a different risk tolerance - choosing the right one depends on the manager's context and personality.
Bayesian Decision Making
A dynamic approach that evolves with new information - moving beyond static models to continuously refine business
decisions.
The Bayesian Framework
Prior Probability Information Updating Application
Use existing knowledge or initial Revise decisions as new market Used in risk analysis and hypothesis
estimates before new data is information arrives - unlike static testing to refine business forecasts.
collected. models.
The Bayesian approach allows a manager to revise their decisions as new market information arrives - turning uncertainty
into a learning process.
Key distinction: Static models give you one answer. Bayesian models give you a better answer every time you gather
new data.
Decision Tree Analysis
A powerful graphical tool that maps decision paths, chance events, and consequences into a single visual framework.
Anatomy of a Decision Tree
Decision Nodes ■
Where the manager must choose a
course of action.
Outcome
Payoff value calculated
Chance Nodes ●
Chance Node
External factors occur Where external factors or "states of
nature" occur.
Decision Node
Manager selects action
Outcomes
The resulting payoffs at each
terminal branch.
Expected Monetary Value (EMV)
The Core Calculation How It Works
By calculating the EMV at each node, you can • Assign probabilities to each chance outcome
mathematically determine the most profitable decision • Multiply probability × payoff for each branch
path. • Sum the weighted values at each decision node
• Choose the path with the highest EMV
The Complete Decision Toolkit
Foundations
Structured vs. unstructured decisions, model building
Uncertainty Rules
Maximax, Maximin, Laplace, Minimax Regret
Bayesian Updating
Prior probabilities refined with new data
Decision Trees
Visual mapping with EMV evaluation
Key Takeaways
1 Replace Intuition with Science 2 Match the Rule to the Mindset
Decision Science applies mathematical logic to Optimistic, pessimistic, neutral, or regret-averse —
management problems. choose accordingly.
3 Update as You Learn 4 Visualize the Path Forward
Bayesian methods let you refine decisions with every Decision trees and EMV calculations reveal the most
new piece of data. profitable course of action.