Modelling and Decision Analysis

 Managerial Approach to Decision Making

Management
Situation

Decisions

Implementation

Payoff

1
 The Modelling Process

ANALYSES
MODEL RESULTS

A I
B N
S SYMBOLIC WORLD T
T E
R R
A P
REAL WORLD R
C
T E
T

MANAGEMENT INTUITION
SITUATION DECISIONS
2
 Models

 Models provide a framework for logical

and consistent analysis
 Models are representations of real
objects or situations
 Models can be presented in various
forms

3
 Models
• Three forms of models are iconic, analog, and
mathematical.
• Iconic Models (physical) are physical replicas (scalar
representations) of real objects.
• Analogue Models are physical in form, but do not
physically resemble the object being modelled.
• Mathematical Models (symbolic) represent real world
problems through a system of mathematical formulas
and expressions based on key assumptions, estimates,
or statistical analyses (we will study how various
mathematical models can be implemented and
analysed on computers using spreadsheet software).
4
 Mathematical Models
 Cost/benefit considerations must be made in selecting an
appropriate mathematical model.
 Frequently a less complicated (and perhaps less precise)
model is more appropriate than a more complex and
accurate one due to cost and ease of solution
considerations.

Example of Mathematical Models

Profit = f(Revenue, Expense)
Or…let Y be Profit, X1, be Revenue and X2 be Expense,
then we can write the function as:
Y = f(X1, X2).
5
 Mathematical Models

 Mathematical models relate decision variables (or

controllable inputs) with fixed or variable parameters (or
uncontrollable inputs).
 Frequently mathematical models seek to maximize or
minimize some objective function subject to constraints.
 If any of the uncontrollable inputs is subject to variation
the model is said to be stochastic; otherwise the model is
said to be deterministic.
 Generally, stochastic models are more difficult to analyse.
 The values of the decision variables that provide the
mathematically-best output are referred to as the optimal
solution for the model.

6
 Benefits of Modeling
• Economy - it is often less costly to
analyze decision problems using models.
• Timeliness - models often deliver needed
information more quickly than their real-
world counterparts.
• Feasibility - models can be used to do
things that would be impossible.
• Models give us insight & understanding
that improves decision making.

7
 Reasons for Modelling Approach

 Potential Reasons for applying the

Modelling Approach to Decision
Making
 The problem is complex.
 The problem is very important.
 The problem is new.
 The problem is repetitive.

8
 Characteristics of Models

• Models are usually simplified

versions of the things they represent
• A valid model accurately represents
the relevant characteristics of the
object or decision being studied

9
 Categories of Mathematical Models
Model Characteristics
Values of Management
Independent Science
Category Form of f() Variable Techniques
Prescriptive Known, Known or under LP, Networks,
Models Well-known decision Maker’s IP, CPM,
control EOQ, NLP

Predictive Unknown, Known or under Regression

Models Ill-defined decision maker’s Analysis, Time
control Series Analysis
Discriminant
Analysis

Descriptive Known, Unknown or Simulation,

Models Well-defined Uncertain Queuing,
Inventory
Models

10
 Problem Solving and Decision
Making (The Process)
7 Steps of Problem Solving
 Identify and define the problem…formulate
problem.
 Data Collection.
 Model Formulation.
 Model Verification
 Evaluate alternatives and choose ...
 Presentation of Results.
 Implement the chosen alternative +
Evaluate the results.

11
 1. Problem Definition

 If organisation (client) believes it has a

problem it commissions a study.
 Analyst learns of the clients cares and
concerns.
 Analyst does his/her own investigations.

12
 2. Data Collection

 Crucial but tedious.

 Client’s data may not be in the form
analyst requires.
 Data may be scattered far and wide
(and in different organisations).
 Analyst must (or may have to):
 Ask questions; do studies; observe people
and other firm entities at work; etc

13
 3. Model Formulation
 Requires special skill of management
scientist
 Focus of this course
 Analyst formulates a mathematical
model of the problem (or more than one
models)
 Model should represent the client’s
problem accurately
 KISS (model should be simple)

14
 4. Model Verification

 Client likely to accept the model if some

type of verification is supplied.
 Verification may include
 Use model with the company’s current values
of the input parameters (outputs should then
be in line with client’s experience)
 Using input parameters (even if not from the
client) to see if outputs are reasonable

15
 Model Debugging (Steps 3 & 4)

Formulate & Analyse Test

Test Model Model Results

unsatisfactory
results

16
 5. Selecting Alternative
 After the first 4 steps, use the model to
make recommendations on decisions or
strategies
 We may seek to optimize an objective
function, but also may find the
alternatives that best achieve multiple
objectives.
 Not all solutions are optimal – problem
may be too complex. Then analyst may
have heuristic solutions. We may have
to choose among heuristic alternatives
17
 6. Presentation of Results
 Sooner or later analyst must present
results to client. There may be a
communication gap between the two
because of technical jargon. Results may
be rejected.
 Key is to include key stakeholders from
the beginning.
 Model should be as intuitive and user
friendly as possible.

18
 7. Implementation of the Model
 Work is not complete until the
recommendations have been
implemented.
 Often this is only possible after step 6
 People who must use results daily must
be fully trained in its use
 Clients like menu-driven systems,
graphics, ability to ask what if.
 Analyst should evaluate results to see
that they are in line with predictions.
19
 Flowchart for Seven-Step Process
Define Possible Feedback Loops
(under- Observe
stand) system,
Problem Collect Form-
data ulate
Math Verify
Model Model, use
Select
for
suitable
prediction
alternative Present
results to
organization

Implement
and
evaluate
 Good Decisions vs. Good Outcomes
• Good decisions do not always lead to good
outcomes...

A structured, modeling approach to decision

making helps us make good decisions, but
can’t guarantee good outcomes.

21