Lesson 7 - The Modeling Process

Question 1

Model

Answer 1

Mathematical, economic, or statistical representation of relationships that determine the probable future consequences of decision choices

Any effort ranging from a back-of-the-envelope simple quantitative analysis to a large, complex computerized system that attempts to describe and/or evaluate alternatives when faced with a decision.

Question 2

Model

Answer 2

The application of mathematical constructs to decision making.

Question 3

Models used…

Answer 3

to assist decision makers to make higher quality business decisions than they would otherwise make if they didn’t use a model.

Question 4

Models can help improve functions…

Answer 4

such as:
 - strategic planning
- manufacturing plans
- marketing plans
- product distribution plans
... many other applications

Question 5

Models used by:

Answer 5

Military - plan and test strategic and tactical plans; determine need for spare parts, conduct military games to assess effectiveness for military hardware.

Hospitals - schedule use of operating rooms; schedule services for support personnel

Emergency Services Units - ambulance, police, firefighters, medical assistance personnel to plan facilities to maximize service and minimize costs

Schools - use location analysis to determine where schools should be located

Question 6

Groups/Types of Models

Answer 6

1. Simple
2. Intermediate
3. Complex

Question 7

Simple Models

Answer 7

When to use/Given:
- a one time calculation is needed to determine the likely values of some business parameters
- the required logic is obvious or easily derived
- ready availability of input data
Requirements:
- perhaps no model is needed - or a simple back-of-the-envelope, quick response effort
-necessary calculations can be done with a hand calculator or a simple spread sheet

Question 8

Intermediate Models

Answer 8

Given:

• an easily identified set of appropriate decision variables and a range of expected values for the decision variables
• generally accepted logic describing relations between decision variables and range of expected values for the decision variables
• Input data are reaily available or can be derived with a reasonable expenditure of effort

Question 9

Intermediate Models - Requirements

Answer 9

• means of specifying values of some decision
• simple what-if model that:
  - contains the logic (set of relations between decision variables)
  - accepts values for the independent decision variables
  - calculates the resultant values of the dependent variables
  - can be created on any computer (PC or mainframe)
  - appropriate including spreadsheet type of software

Question 10

Complex Models

Answer 10

Given:
- the client has a number of quantifiable decision alternatives
- reasonable expectation that decision variables can be identified
reasonable expectation that the required input data can be developed, albeit with a large effort

Question 11

Complex Models - Requirements

Answer 11

Requirements:

• powerful but flexible model is required to simulate the probable outcomes of various alternatives
• a means of specifying a base case against which other alternatives are compared
• tool that enables the client to investigate trade-offs between various decision variables
• powerful software modeling tools are required to accomplish the effort in a reasonable timeframe
  ** May include:
  GPSS (simulation language);
  General Purpose Modeling Language - APL, LISP, FORTRAN, Basic, C, C+, C++, C#, Java, Pascal, VB
• programming shells (specialized software tools that an be used to develop quite complex models)
• High powered PC. workstation, mainframe

Question 12

Modeling Tools

Answer 12

SIMPLEST - paper, pencil, handheld calculator

NEXT - Spreadsheets -Excel, Lotus, Quattro Pro –> user friendly, easy to learn, develop wide range of models using SS

Programming Languages - general programming languages - develop simple one-time calculations to large computer-based simulations of complex systems

Programming Tools - developed for specific applications
- tools enable you to solve complex business problems through application of Management Science principles/techniques

Question 13

Two Modeling Approaches

Answer 13

1. Reductionist

2. Holistic

Question 14

Reductionist

Answer 14

break things down into their component parts and study each of them - once the parts and sub-parts are understood the system as a whole will be understood
–> EFFECTIVE MEANS TO UNDERSTAND HOW A SYSTEM WORKS

Question 15

Holistic Approach

Answer 15

starts with entire system and studies that system in terms of how it interacts with its environment under various conditions

Question 16

Requirements for Successful Modeling

Answer 16

1. Be a domain expert
2. know enough about model development that you have skills to develop model

** must understand system under study (the cause and effect relationship) and also know how to use analytical techniques

Question 17

Modeling Truisms

Answer 17

1. Modeling is part science, part art.
2. Judgment supported by analysis is always more effective than judgment alone.
3. Complex decisions require sophisticated analysis and take longer to make than a simpler decision.
4. Quick and dirty analyses are better than gut feeling/instinct. - Even the simplest, quickest analysis provides some insight that leads to a better decision.
5. Rigorous analyses are always better.
6. Be conservative - err on the side of caution.
7. Don’t marry a particular approach. Be proficient in a variety of approaches and apply the most appropriate tool for the situation.
8. Good analytical approaches are timeless.
9. Avoid extremes: seat-of-the pants decision making and analysis paralysis.

Question 18

Quick & Dirty Approach

Answer 18

A not so elegant method that gives decision maker something to go on in a short time frame and with a minimum effort. Advantage: gives a ballpark to tell you if your idea is feasible or not.