MANAGEMENT INFORMATION SYSTEMS 8/E

Raymond McLeod, Jr. and George Schell

Chapter 13
Decision Support Systems

13-1
Copyright 2001 Prentice-Hall, Inc.
 Objectives
Q Have an expanded theoretical base for understanding
decision making and the DSS concept.
Q Understand the objectives of a DSS.
Q Understand one definition of a DSS and its accompanying
model.
Q Know how to apply the DSS concept to group problem
solving and achieve a group decision support system/GDSS
Q Be familiar with special GDSS software called groupware.
Q Know what is meant by the term artificial intelligence as
well as what areas are included.
Q Understand the appeal of expert sistem and how they
compare with DSSs.
 Simons Types of Decisions

Q Programmed decisions
repetitive and routine
have a definite procedure
Q Nonprogrammed decisions
Novel and unstructured
No cut-and-dried method for handling problem
Q Types exist on a continuum

13-3
Simons Problem Solving Phases
Q Intelligence
Searching environment for conditions calling for a
solution
Q Design
Inventing, developing, and analyzing possible courses
of action
Q Choice
Selecting a course of action from those available
Q Review
Assessing past choices
13-4
 Definitions of a Decision
Support System (DSS)
General definition - a system providing both
problem-solving and communications capabilities
for semistructured problems
Specific definition - a system that supports a
single manager or a relatively small group of
managers working as a problem-solving team in
the solution of a semistructured problem by
providing information or making suggestions
concerning specific decisions.
13-5
 The DSS Concept
Q Gorry and Scott Morton coined the phrase
DSS in 1971, about ten years after MIS
became popular
Q Decision types in terms of problem structure
Structured problems can be solved with
algorithms and decision rules
Unstructured problems have no structure in
Simons phases
Semistructured problems have structured and
unstructured phases 13-6
 The Gorry and Scott Morton Grid
Management levels
Operational Management Strategic
control control planning
Structured Accounts Budget analysis-- Tanker fleet
receivable engineered costs mix
Degree of Order entry Short-term Warehouse and
problem forecasting factory location
structure Inventory
control

Semistructured Production Variance analysis-- Mergers and

scheduling overall budget acquisitions
Cash Budget
management New product
preparation planning
PERT/COST Sales and
Unstructured systems R&D planning
production

13-7
 Alters DSS Types
Q In 1976 Steven Alter, a doctoral student
built on Gorry and Scott-Morton framework
Created a taxonomy of six DSS types
Based on a study of 56 DSSs
Q Classifies DSSs based on degree of
problem solving support.

13-8
 Levels of Alters DSSs
Q Level of problem-solving support from
lowest to highest
Retrieval of information elements
Retrieval of information files
Creation of reports from multiple files
Estimation of decision consequences
Propose decisions
Make decisions

13-9
 Importance of Alters Study

Q Supports concept of developing systems

that address particular decisions
Q Makes clear that DSSs need not be
restricted to a particular application type

13-10
 Alters DSS Types

Degree
of
problem
Retrieve Analyze Prepare Estimate Propose Make solving
information entire reports decision decisions decisions support
elements files from consequen-
multiple ces
files

Degree of
Little Much
complexity of the
problem-solving
system 13-11
 Three DSS Objectives
1. Assist in solving semistructured problems
2. Support, not replace, the manager
3. Contribute to decision effectiveness, rather
than efficiency

Based on studies of Keen and Scott-Morton

13-12
A DSS Model
Environment
Individual Other
problem group
solvers members

Report GDSS
Mathematical GDSS
writing software
Models software
software

Database

Decision
support
system
Environment
Data Communication Information
Legend:
13-13
 Database Contents
Q Used by Three Software Subsystems
Report writers
Special reports
Periodic reports
COBOL or PL/I
DBMS
Mathematical models
Simulations
Special modeling languages
Groupware or GDSS
13-14
Group Decision Support Systems
Q Computer-based system that supports
groups of people engaged in a common task
(or goal) and that provides an interface to a
shared environment.
Q Used in problem solving
Q Related areas
Electronic meeting system (EMS)
Computer-supported cooperative work (CSCW)
Group support system (GSS)
Groupware 13-15
 How GDSS Contributes
to Problem Solving
Q Improved communications
Q Improved discussion focus
Q Less wasted time

13-16
 GDSS Environmental Settings
Q Synchronous exchange
Members meet at same time
Committee meeting is an example
Q Asynchronous exchange
Members meet at different times
E-mail is an example
Q More balanced participation.
13-17
 GDSS Types
Q Decision rooms
Small groups face-to-face
Parallel communication
Anonymity
Q Local area decision network
Members interact using a LAN
Q Legislative session
Large group interaction
Q Computer-mediated conference
Permits large, geographically dispersed group
interaction 13-18
Group Size and Location Determine
GDSS Environmental Settings
GROUP SIZE
Smaller Larger

Face-to- Decision Legislative

face
Room Session
MEMBER
PROXIMITY Local Area Computer-
Dispersed Decision Mediated
Network Conference

13-19
 Groupware
Q Functions
E-mail
FAX
Voice messaging
Internet access
Q Lotus Notes
Popular groupware product
Handles data important to managers
13-20
 Main Groupware Functions
IBM TeamWARE Lotus Novell
Function Workgroup Office Notes GroupWise
E lectro n ic m ail X X X X
F AX X X O X
V o ice m essag in g O X
In tern et access X X O X
Bu lletin b o ard system X 3 O
P erso n al calen d arin g X X 3 X
G ro u p calen d arin g X X O X
E lectro n ic co n feren cin g O X 3 3
T ask m an ag em en t X X 3 X
Deskto p vid eo co n feren O
Datab ase access O X 3
W o rkflo w ro u tin g O X 3 X
Reen g in eerin g O X 3
E lectro n ic fo rm s O 3 3 O
G ro u p d o cu m en ts O X X O
X = standard feature O = optional feature 3 = third party offering
13-21
Artificial Intelligence (AI)

The activity of providing such

machines as computers with the
ability to display behavior that
would be regarded as intelligent if
it were observed in humans.

13-22
 History of AI

Q Early history
John McCarthy coined term, AI, in 1956, at
Dartmouth College conference.
Logic Theorist (first AI program. Herbert Simon
played a part)
General problem solver (GPS)
Q Past 2 decades
Research has taken a back seat to MIS and DSS
development
13-23
Areas of Artificial Intelligence

Expert
systems AI
hardware
Robotics
Natural
language Perceptive
systems
Neural (vision,
networks hearing)
Learning

Artificial Intelligence
13-24
 Appeal of Expert Systems
Q Computer program that codes the
knowledge of human experts in the form of
heuristics
Q Two distinctions from DSS
1. Has potential to extend managers problem-
solving ability
2. Ability to explain how solution was reached

13-25
 Instructions & Solutions &
Knowledge
User information explanations

User
interface

Know-
Inference Problem
ledge
engine Domain
base

Development
engine
Expert
system
An Expert
System Model
Expert and
knowledge engineer 13-26
 Expert System Model
Q User interface
Allows user to interact with system
Q Knowledge base
Houses accumulated knowledge
Q Inference engine
Provides reasoning
Interprets knowledge base
Q Development engine
Creates expert system
13-27
 User Interface

Q User enters:
Instructions
Information } Menus, commands, natural language, GUI

Q Expert system provides:

Solutions
Explanations of
Questions
Problem solutions

13-28
 Knowledge Base

Q Description of problem domain

Q Rules
Knowledge representation technique
IF:THEN logic
Networks of rules
Lowest levels provide evidence
Top levels produce 1 or more conclusions
Conclusion is called a Goal variable.

13-29
 A Rule Set That
Conclusion Produces One Final
Conclusion

Conclusion
Conclusion

Evidence Evidence Evidence Evidence

Evidence Evidence Evidence Evidence

13-30
 Rule Selection

Q Selecting rules to efficiently solve a

problem is difficult
Q Some goals can be reached with only a few
rules; rules 3 and 4 identify bird

13-31
 Inference Engine

Q Performs reasoning by using the contents of

knowledge base in a particular sequence
Q Two basic approaches to using rules
1. Forward reasoning (data driven)
2. Reverse reasoning (goal driven)

13-32
 Forward Reasoning
(Forward Chaining)
Q Rule is evaluated as:
(1) true, (2) false, (3) unknown
Q Rule evaluation is an iterative process
Q When no more rules can fire, the reasoning
process stops even if a goal has not been
reached
Start with inputs and
work to solution
13-33
Rule 1 The Forward
T
IF A
THEN B Rule 7 Reasoning
Rule 2
F
IF B OR D
THEN K
Rule 10
IF K AND
Process
IF C T L THEN N
THEN D
T
Rule 3 Rule 8 Rule 12
T T
IF M IF E IF N OR O
THEN E THEN L THEN P
T
Rule 4
T
IF K Legend:
THEN F
Rule 9 First pass
Rule 5 Rule 11
T IF T
IF G IF(F
(FAND
ANDH)
H) IF M
OR
ORJJ
THEN H THEN O Second pass
THEN
THENMM T
Rule 6
F Third pass
IF I
THEN J
13-34
 Reverse Reasoning Steps
(Backward Chaining)
Divide problem into subproblems
Try to solve one subproblem
Then try another

Start with solution

and work back to
inputs

13-35
 Step 4 The First Five Problems
Rule 1
IF A THEN
Step 3 Are Identified
B Rule 7
T Step 2
IF B OR D Rule 10
THEN K Step 1
Rule 2 T IF K AND L Rule 12
IF C THEN N
IF N OR O
THEN D
THEN P
Step 5
Rule 3 Rule 8
IF M IF E
THEN E THEN L

Rule 11 Legend:
Problems to
IF (F AND H) IF M
M be solved
Rule 9 IF
OR J THEN O
THEN O
THEN M
13-36
 The Next Four Problems Are
Step 8 Identified Rule 12

Rule 4
If N Or O
Then P T
If K
Then F
T
Step 7 Step 6
Step 9
Rule 5
If G IF (F And H) If M
Then H Or J Then O
T Then M T T

Rule 9 Rule 11 Legend:

Rule 6 Problems to
be solved
If I
Then J 13-37
Forward Versus Reverse Reasoning
Q Reverse reasoning is faster than forward
reasoning
Q Reverse reasoning works best under certain
conditions
Multiple goal variables
Many rules
All or most rules do not have to be examined in
the process of reaching a solution
13-38
 Development Engine
Q Programming languages
Lisp
Prolog
Q Expert system shells
Ready made processor that can be tailored to a
particular problem domain
Q Case-based reasoning (CBR)
Q Decision tree

13-39
 Expert System Advantages

Q For managers
Consider more alternatives
Apply high level of logic
Have more time to evaluate decision rules
Consistent logic
Q For the firm
Better performance from management team
Retain firms knowledge resource
13-40
 Expert System Disadvantages

Q Cant handle inconsistent knowledge

Q Cant apply judgment or intuition

13-41
 Keys to Successful ES
Development
Q Coordinate ES development with strategic
planning
Q Clearly define problem to be solved and
understand problem domain
Q Pay particular attention to ethical and legal
feasibility of proposed system
Q Understand users concerns and expectations
concerning system
Q Employ management techniques designed to retain
developers 13-42
 Neural Networks

Q Mathematical model of the human brain

Simulates the way neurons interact to process
data and learn from experience
Q Bottom-up approach to modeling human
intuition

13-43
 The Human Brain

Q Neuron -- the information processor

Input -- dendrites
Processing -- soma
Output -- axon
Q Neurons are connected by the synapse

13-44
Simple Biological Neurons

Soma
Axonal Paths
(processor)
(output)

Synapse
Axon

Dendrites
(input)

13-45
 Evolution of Artificial
Neural Systems (ANS)
Q McCulloch-Pitts mathematical neuron
function (late 1930s) was the starting point
Q Hebbs learning law (early 1940s)
Q Neurocomputers
Marvin Minskys Snark (early 1950s)
Rosenblatts Perceptron (mid 1950s)

13-46
 Current Methodology

Q Mathematical models dont duplicate

human brains, but exhibit similar abilities
Q Complex networks
Q Repetitious training
ANS learns by example

13-47
 Single Artificial Neuron
y1 w1

y2 w2

y3 w3

y
wn-1

yn-1 13-48
 OUT1 OUTn

Input
The Multi-Layer
Layer Perceptron

Y1 Yn2

OutputL
ayer

IN1 INn
13-49
 Knowledge-based Systems
in Perspective
Q Much has been accomplished in neural nets
and expert systems
Q Much work remains
Q Systems abilities to mimic human
intelligence are too limited and regarded as
primitive

13-50
 Summary [cont.]

Q AI
Neural networks
Expert systems
Q Limitations and promise

13-51
 Case Study
1. A decision support system should let a manager
see the possible effects of a decision.
A) true B) false

2. One type of decision support system (identified

by Steve Alter) can make decisions for a
manager.
A) true B) false