An Introduction to Management

Science
Quantitative Approaches to Decision Making

Chapter 1: Introduction
1.1. Problem Solving and Decision Making
7 Steps in Problem Solving
(First 5 are the process of Decision Making)
1. Define the problem
2. Determine the set of alternative solutions.
- Laying out the options
3. Determine the criteria for evaluating alternatives.
4. Evaluate the alternatives. Using the criteria
5. Choose an alternative (make a decision).
6. Implement the selected alternative.
7. Evaluate the results.
1.2. Quantitative Analysis and Decision Making
Decision-Making Process
a. Single-criterion decision problems
- Only 1 factor is to be considered
b. Multicriteria decision problems
- More than 1 factors are to be considered
Analysis Phase of Decision-Making Process
Qualitative Analysis
- Based on manager’s judgement and experience
- Manager’s intuitive “feel” for the problem
- More of an art than a science

Quantitiative Analysis
- Based on quantitative facts or data associated w/ the problem
- Mathematical expressions (objectives, constrainst, and other relationships
that exist in the problem)
- Uses one or more quantitative methods to make a recommendation
- For complex, very important, new, and repetitive problems
- Process:
● Model Development
- Requires less time
- Less expensive
- Less risk
3 forms of model:
● Iconic models → physical replicas of real objects
● Analog models → physical in form, but do not physically resemble the
object being modeled
 ● Mathematical models → a system of mathematical formulas and
expressions that represent real world problems based on
assumptions, estimates, or statistical analyses
○ Objective Function - mathematical expression describing the
problem’s maximizing profit / minimizing cost
○ Constraints - set of restrictions or llimitations
○ Uncontrollable Inputs - environmental factors that are not
under the control of the decision maker
○ Decision Variables - controllable inputs, ex. Number of units
of a product to produce

● Data Preparation
● Model Solution
Optimal Solution - best output for the model
Infeasible - alternative does not satisfy all of the model constraints
Feasible - alternative satisfies all of the model constraints and is a candidate
for the “best” solution

● Report Generation
1.3. Models of Cost, Revenue, and Profit
● Cost
○ Fixed cost → does not depend on production volume
○ Variable cost → varies with the production volume
● Marginal Cost
● Marginal Revenue
● Breakeven Point
1.4. Management Science Techniques
Most used:
● Linear programming
● Integer linear programming
● Network models
● Simulation
 ● Project scheduling: PERT (Program Evaluation and Review Technique) and
CPM (Critical Path Method)
● Inventory Models
● Waiting line (or queuing) models
● Decision analysis
● Forecasting methods
● Goal programming
● Analytic hierarchy process
● Markov-process models

CHAPTER 13: Decision Analysis

● Decision analysis
- used to develop an optimal strategy when a decision maker is faced
with several decision alternatives and an uncertain or risk-filled pattern
of future events

○ Risk analysis
- Provides probability information about the favorable and the
unfavorable consequences that may occur
A decision problem is characterized by:
a. Decision alternatives
- The different possible strategies the decision maker can employ
b. States of nature
- Future events, not under the control of the decision maker, which may
occur.
c. Resulting payoffs

● Influence diagram
- Graphical device showing the relationships among the decisions, the chance
events, and the consequences
○ Squares or rectangles → decision nodes
○ Circles and ovals → chance nodes
○ Diamonds → consequence nodes
○ Lines or arcs connecting the nodes show the direction of influence

● Payoff tables
○ Payoff - consequence resulting from a specific combination of a decision
alternative and a state of nature
○ Payoff table - table showing payoffs for all combinations of a decision
alternative and a state of nature
○ Profit, cost, time, distance, etc.

Decision Making w/o Probabilities

 ● Optimistic approach (Minimin)
- Largest payoff or lowest cost is chosen
● Conservative approach (maximin)
- Minimum payoff is listed for each decision, maximum payoff is selected
- Maximum costs are determined and the minimum of those is selected
- Maximum of thee minimum payoffs
- Minimum of the maximum costs

● Minimax regret approach

- Construction of regret or opportunity loss table
- Difference between each payoff and the largest payoff for that state of nature
- Maximum regret for each possible decision is listed
- Decision chosen is the minimum of the maximum regrets

Decision Making with Probabilities

● Expected value approach
- Expected return = (Products of the payoff under each state of nature) +
(probability of the respective state of nature occuring)
- Decision that yields the best expected return is chosen
- Sum the products of each payoff and the probability of its state of nature
occurring

● Expected value of a decision alternative (EV)

- Sum of weighted payoffs for the decision alternative
EV(d1) = () + ()
● Decision tree
- Chronological representation of the decision problem
○ Square nodes → decision alternatives
○ Round nodes → states of nature
○ End of each limb → payoffs

● Expected Value of Perfect Information (EVPI)

- Increase in the expected profit if one knew with certainty which state of nature
would occur
- Provides an upper bound on the expected value of any sample or survey
information
EVPI = |EVwPI - EVwoPI|
Expected Value with Perfect Information - Expected Value without Perfect Information

● Risk analysis
- To recognize the difference between
a. EV of a decision alternative
b. Payoff that might actually occur
○ Risk profile
- Shows possible payoffs along with their associated probabilities
 ● Sensitivity analysis
- Used to determine how changes to the following inputs affect the
recommended decision alternative:
a. Probabilities for the states of nature
b. Values of the payoffs

Decision Analysis with Sample Information

● Posterior probabilities
- Revised probabilities of prior probabilities

● Sample information
If market research is undertaken:
P(Favorable report)
P(Ufavourable report)
If market research report is favorable:
P(Strong demand | F) = P(S1 | F)
P(Weak demand| F) = P(S2 | F)
If market research is unfavourable
P(Strong demand | U)
P(Weak demand| U)
If market research is not undertaken, prior probabilities are applicable
P(Favorable report)
P(Ufavourable report)

● Decision strategy
- A sequence of decisions and chance outcomes
- Determined based on a backward pass through the decision tree:
■ Chance nodes = compute EV by multiplying payoff and probabilities
■ Decision nodes = select decision branch with best EV

● Expected Value of Sample Information (EVSI)

- Additional expected profit possible through knowledge of the sample or
survey information
- EV with Market research - best EV w/o market research = EVSI

● Efficiency of sample information

- Ratio of EVSI to EVPI
- Always a number between 0 and 1

E = (EVSI/EVPI) X 100

● Computing Branch Probabilities

1. For each state of nature, multiply the prior probability by its condition
probability → joint probabilities
2. Sum these joint probabilities over all states → marginal probability
3. Divide each joint probabiltiy by the marginal probability → Posterior
probability distribution
● Baye’s Theorem and Posterior Posibilities
Prior probabilities → obtaining conditional probabilities of sample information →
Posterior probabilities or branch probabilities for decision trees

sj → P(s) → P (F | sj) → P(F | sj) → P (sj | F)

sj → P(s) → P (U | sj) → P(U | sj) → P (sj | U)