Chapter 13Decision Support SystemsDecision Support Systems

MANAGEMENT INFORMATION SYSTEMS 8/ERaymond McLeod, Jr. and George Schell

Copyright 2001 Prentice-Hall, Inc.

13-1

ObjectivesObjectivesHave an expanded theoretical base for understandingdecision making and the DSS concept.Understand the objectives of a DSS.Understand one definition of a DSS and its accompanyingmodel.Know how to apply the DSS concept to group problemsolving and achieve a group decision support system/GDSSBe familiar with special GDSS software called groupware.Know what is meant by the term artificial intelligence aswell as what areas are included.Understand the appeal of expert sistem and how theycompare with DSSs.

Simon’s Simon’s Types of DecisionsTypes of Decisions

Programmed decisionsProgrammed decisions–– repetitive and routinerepetitive and routine–– have a definite procedurehave a definite procedure

Nonprogrammed decisionsNonprogrammed decisions–– Novel and unstructuredNovel and unstructured–– No cut-and-dried method for handling problemNo cut-and-dried method for handling problem

Types exist on a continuumTypes exist on a continuum

13-3

Simon’sSimon’s Problem Solving Phases Problem Solving Phases

IntelligenceIntelligence–– Searching environment for conditions calling for aSearching environment for conditions calling for a

solutionsolution

DesignDesign–– Inventing, developing, and analyzing possible coursesInventing, developing, and analyzing possible courses

of actionof action

ChoiceChoice–– Selecting a course of action from those availableSelecting a course of action from those available

ReviewReview–– Assessing past choicesAssessing past choices

13-4

Definitions of a DecisionDefinitions of a DecisionSupport System (DSS)Support System (DSS)

General definition - General definition - a system providing botha system providing bothproblem-solving and communications capabilitiesproblem-solving and communications capabilitiesfor semistructured problemsfor semistructured problems

Specific definition - Specific definition - a system that supports aa system that supports asingle manager or a relatively small group ofsingle manager or a relatively small group ofmanagers working as a problem-solving team inmanagers working as a problem-solving team inthe solution of a semistructured problem bythe solution of a semistructured problem byproviding information or making suggestionsproviding information or making suggestionsconcerning concerning specific specific decisions.decisions.

13-5

The DSS ConceptThe DSS Concept

GorryGorry and Scott Morton coined the phrase and Scott Morton coined the phrase‘DSS’ in 1971, about ten years after MIS‘DSS’ in 1971, about ten years after MISbecame popularbecame popularDecision types in terms of problem structureDecision types in terms of problem structure–– Structured problems can be solved withStructured problems can be solved with

algorithms and decision rulesalgorithms and decision rules–– Unstructured problems have no structure inUnstructured problems have no structure in

Simon’sSimon’s phases phases–– Semistructured problems have structured andSemistructured problems have structured and

unstructured phasesunstructured phases13-6

Degree ofDegree ofproblemproblemstructurestructure

TheThe Gorry Gorry and Scott Morton Grid and Scott Morton GridManagement levelsManagement levels

Structured Structured

SemistructuredSemistructured

Unstructured Unstructured

OperationalOperational control control

ManagementManagement control control

StrategicStrategicplanningplanning

Accountsreceivable

Order entry

Inventory control

Budget analysis--engineered costs

Short-term forecasting

Tanker fleet mix

Warehouse andfactory location

Productionscheduling

Cashmanagement

PERT/COST systems

Variance analysis-- overall budget

Budget preparation

Sales and production

Mergers and acquisitions

New product planning

R&D planning

13-7

Alter’s DSS TypesAlter’s DSS Types

In 1976 Steven Alter, a doctoral studentIn 1976 Steven Alter, a doctoral studentbuilt on built on Gorry Gorry and Scott-Morton frameworkand Scott-Morton framework–– Created a taxonomy of six DSS typesCreated a taxonomy of six DSS types–– Based on a study of 56 DSSsBased on a study of 56 DSSs

Classifies DSSs based on “degree ofClassifies DSSs based on “degree ofproblem solving support.”problem solving support.”

13-8

Levels of Alter’s DSSsLevels of Alter’s DSSs

Level of problem-solving support fromLevel of problem-solving support fromlowest to highestlowest to highest–– Retrieval of information elementsRetrieval of information elements–– Retrieval of information filesRetrieval of information files–– Creation of reports from multiple filesCreation of reports from multiple files–– Estimation of decision consequencesEstimation of decision consequences–– Propose decisionsPropose decisions–– Make decisionsMake decisions

13-9

Importance of Alter’s StudyImportance of Alter’s Study

Supports concept of developing systemsSupports concept of developing systemsthat address particular decisionsthat address particular decisionsMakes clear that DSSs need not beMakes clear that DSSs need not berestricted to a particular application typerestricted to a particular application type

13-10

RetrieveRetrieveinformationinformation

elementselements

AnalyzeAnalyzeentireentirefilesfiles

PreparePreparereportsreports

fromfrommultiplemultiple

filesfiles

EstimateEstimatedecisiondecision

consequenconsequen--cesces

ProposeProposedecisionsdecisions

DegreeDegreeofofproblemproblemsolvingsolvingsupportsupport

Degree ofDegree ofcomplexity of thecomplexity of theproblem-solvingproblem-solving

systemsystem

LittleLittle MuchMuch

Alter’s DSS TypesAlter’s DSS Types

MakeMakedecisionsdecisions

13-11

Three DSS ObjectivesThree DSS Objectives

1.1. Assist in solving semistructured problems Assist in solving semistructured problems2.2. Support, not replace, the manager Support, not replace, the manager3.3. Contribute to decision effectiveness, rather Contribute to decision effectiveness, rather

than efficiencythan efficiency

Based on studies of Keen and Scott-Morton

13-12

GDSS software

MathematicalMathematicalModelsModels

OtherOther group group

membersmembers

DatabaseDatabase

GDSSGDSSsoftwaresoftware

EnvironmentEnvironment IndividualIndividual problem problem solvers solvers

Decision support system

EnvironmentEnvironment

Legend:Data Information Communication

A DSS ModelA DSS Model

ReportReportwritingwriting

softwaresoftware

13-13

Database ContentsDatabase Contents

Used by Three Software SubsystemsUsed by Three Software Subsystems–– Report writersReport writers

»» Special reportsSpecial reports»» Periodic reportsPeriodic reports»» COBOL or PL/ICOBOL or PL/I»» DBMSDBMS

–– Mathematical modelsMathematical models»» SimulationsSimulations»» Special modeling languagesSpecial modeling languages

–– Groupware or GDSSGroupware or GDSS13-14

Group Decision Support SystemsGroup Decision Support SystemsComputer-based system that supportsComputer-based system that supportsgroups of people engaged in a common taskgroups of people engaged in a common task(or goal) and that provides an interface to a(or goal) and that provides an interface to ashared environment.shared environment.Used in problem solvingUsed in problem solvingRelated areasRelated areas–– Electronic meeting system (EMS)Electronic meeting system (EMS)–– Computer-supported cooperative work (CSCW)Computer-supported cooperative work (CSCW)–– Group support system (GSS)Group support system (GSS)–– GroupwareGroupware

13-15

How GDSS ContributesHow GDSS Contributesto Problem Solvingto Problem Solving

Improved communicationsImproved communicationsImproved discussion focusImproved discussion focusLess wasted timeLess wasted time

13-16

GDSS Environmental SettingsGDSS Environmental Settings

Synchronous exchangeSynchronous exchange–– Members meet at same timeMembers meet at same time–– Committee meeting is an exampleCommittee meeting is an example

Asynchronous exchangeAsynchronous exchange–– Members meet at different timesMembers meet at different times–– E-mail is an exampleE-mail is an example

More balanced participation.More balanced participation.13-17

GDSS TypesGDSS TypesDecision roomsDecision rooms–– Small groups face-to-faceSmall groups face-to-face–– Parallel communicationParallel communication–– AnonymityAnonymity

Local area decision networkLocal area decision network–– Members interact using a LANMembers interact using a LAN

Legislative sessionLegislative session–– Large group interactionLarge group interaction

Computer-mediated conferenceComputer-mediated conference–– Permits large, geographically dispersed groupPermits large, geographically dispersed group

interactioninteraction 13-18

Smaller LargerGROUPGROUP SIZESIZE

Face-to-face

Dispersed

DecisionRoom

Local Area Decision Network

Legislative Session

Computer-Mediated

Conference

MEMBERMEMBERPROXIMITYPROXIMITY

Group Size and Location Determine Group Size and Location Determine GDSS Environmental SettingsGDSS Environmental Settings

13-19

GroupwareGroupware

FunctionsFunctions–– E-mailE-mail–– FAXFAX–– Voice messagingVoice messaging–– Internet accessInternet access

Lotus NotesLotus Notes–– Popular groupware productPopular groupware product–– Handles data important to managersHandles data important to managers

13-20

Electronic mail X X X XFAX X X O XVoice messaging O XInternet access X X O XBulletin board system X 3 OPersonal calendaring X X 3 XGroup calendaring X X O XElectronic conferencing O X 3 3Task management X X 3 XDesktop video conferen ODatabase access O X 3W orkflow routing O X 3 XReengineering O X 3Electronic forms O 3 3 OGroup documents O X X O

Main Groupware FunctionsMain Groupware Functions IBM IBM TeamWARE TeamWARE Lotus Novell Lotus Novell Function Workgroup Office NotesFunction Workgroup Office Notes GroupWise GroupWise

X = standard featureX = standard feature O = optional featureO = optional feature 3 = third party offering3 = third party offering13-21

Artificial Intelligence (AI)Artificial Intelligence (AI)

The activity of providing suchmachines as computers with theability to display behavior thatwould be regarded as intelligent ifit were observed in humans.

13-22

History of AIHistory of AI

Early historyEarly history–– JohnJohn McCarthy McCarthy coined term, AI, in 1956, at coined term, AI, in 1956, at

Dartmouth College conference.Dartmouth College conference.–– Logic Theorist (first AI program. Herbert SimonLogic Theorist (first AI program. Herbert Simon

played a part)played a part)–– General problem solver (General problem solver (GPSGPS))

Past 2 decadesPast 2 decades–– Research has taken a back seat to MIS and DSSResearch has taken a back seat to MIS and DSS

developmentdevelopment13-23

Areas of Artificial IntelligenceAreas of Artificial Intelligence

ExpertExpertsystemssystems AIAI

hardwarehardware RoboticsRobotics

PerceptivePerceptive systems systems (vision, (vision, hearing) hearing)

NeuralNeuralnetworksnetworks

NaturalNatural language language

Learning

Artificial IntelligenceArtificial Intelligence13-24

Appeal of Expert SystemsAppeal of Expert Systems

Computer program that codes theComputer program that codes theknowledge of human experts in the form ofknowledge of human experts in the form ofheuristicsheuristicsTwo distinctions from DSSTwo distinctions from DSS–– 1. Has potential to extend manager’s problem-1. Has potential to extend manager’s problem-

solving abilitysolving ability–– 2. Ability to explain how solution was reached2. Ability to explain how solution was reached

13-25

Know-ledgebase

User

Userinterface

Instructions &information

Solutions &explanations Knowledge

Inference engine

Problem Domain

Expert andknowledge engineer

DevelopmentengineExpertExpert

systemsystem An Expert An Expert System ModelSystem Model

13-26

Expert System ModelExpert System ModelUser interfaceUser interface–– Allows user to interact with systemAllows user to interact with system

Knowledge baseKnowledge base–– Houses accumulated knowledgeHouses accumulated knowledge

Inference engineInference engine–– Provides reasoningProvides reasoning–– Interprets knowledge baseInterprets knowledge base

Development engineDevelopment engine–– Creates expert systemCreates expert system

13-27

User InterfaceUser Interface

User enters:User enters:–– InstructionsInstructions–– InformationInformation

Expert system provides:Expert system provides:–– SolutionsSolutions–– Explanations ofExplanations of

»» QuestionsQuestions»» Problem solutionsProblem solutions

}Menus, commands, natural language, GUI

13-28

Knowledge BaseKnowledge Base

Description of problem domainDescription of problem domainRulesRules–– Knowledge representation techniqueKnowledge representation technique–– ‘IF:THEN’ logic‘IF:THEN’ logic–– Networks of rulesNetworks of rules

»» Lowest levels provide evidenceLowest levels provide evidence»» Top levels produce 1 or more conclusionsTop levels produce 1 or more conclusions»» Conclusion is called a Goal variable.Conclusion is called a Goal variable.

13-29

Evidence

Conclusion

Conclusion

Evidence Evidence Evidence Evidence

Evidence Evidence Evidence

Conclusion

A Rule Set That Produces One Final

Conclusion

13-30

Rule SelectionRule Selection

Selecting rules to efficiently solve aSelecting rules to efficiently solve aproblem is difficultproblem is difficultSome goals can be reached with only a fewSome goals can be reached with only a fewrules; rules 3 and 4 identify birdrules; rules 3 and 4 identify bird

13-31

Inference EngineInference Engine

Performs reasoning by using the contents ofPerforms reasoning by using the contents ofknowledge base in a particular sequenceknowledge base in a particular sequenceTwo basic approaches to using rulesTwo basic approaches to using rules–– 1. Forward reasoning (data driven)1. Forward reasoning (data driven)–– 2. Reverse reasoning (goal driven)2. Reverse reasoning (goal driven)

13-32

Forward ReasoningForward Reasoning(Forward Chaining)(Forward Chaining)

Rule is evaluated as:Rule is evaluated as:–– (1) true, (2) false, (3) unknown(1) true, (2) false, (3) unknown

Rule evaluation is an iterative processRule evaluation is an iterative processWhen no more rules can fire, the reasoningWhen no more rules can fire, the reasoningprocess stops even if a goal has not beenprocess stops even if a goal has not beenreachedreached

Start with inputs andwork to solution

13-33

Rule 1Rule 1

Rule 3Rule 3

Rule 2Rule 2

Rule 4Rule 4

Rule 5 Rule 5

Rule 6 Rule 6

Rule 7Rule 7

Rule 8 Rule 8

Rule 9Rule 9

Rule 10Rule 10

Rule 11 Rule 11

Rule 12Rule 12

IF ATHEN B

IF CTHEN D

IF MTHEN E

IF KTHEN F

IF GTHEN H

IF ITHEN J

IF B OR DTHEN K

IF ETHEN L

IF K AND L THEN N

IF M THEN O

IF N OR OTHEN P

F

IF (F AND H)OR JTHEN M

IF (F AND H)OR JTHEN M

The ForwardThe ForwardReasoningReasoning

ProcessProcess

T

TT

T

T

T

T

T

T

F

T

Legend:Legend: First pass

Second pass

Third pass

13-34

Reverse Reasoning StepsReverse Reasoning Steps(Backward Chaining)(Backward Chaining)

Divide problem into subproblemsDivide problem into subproblemsTry to solve one Try to solve one subproblemsubproblemThen try anotherThen try another

Start with solutionand work back toinputs

13-35

T

Rule 1

Rule 2

Rule 3

Rule 9

Rule 11 Legend:Problems to be solved

Step 4

Step 3

Step 2

Step 1

Step 5

IF A THENB

IF B OR DTHEN K

IF K AND LTHEN N

IF N OR O THEN P

IF CTHEN D

IF MTHEN E

IF ETHEN L

IF (F AND H)OR JTHEN M

IF MTHEN O

IF MTHEN O

T

The First Five ProblemsThe First Five ProblemsAre IdentifiedAre Identified

Rule 7

Rule 10

Rule 12

Rule 8

13-36

If KThen F

Legend:Problems to be solved

If GThen H

If IThen J

If MThen O

Step 8

Step 9Step 7 Step 6

Rule 4

Rule 5

Rule 11Rule 6

T

IF (F And H)Or J

Then MT

Rule 9

T T

Rule 12

T

If N Or OThen P

The Next Four Problems AreThe Next Four Problems AreIdentifiedIdentified

13-37

Forward Versus Reverse ReasoningForward Versus Reverse Reasoning

Reverse reasoning is faster than forwardReverse reasoning is faster than forwardreasoningreasoningReverse reasoning works best under certainReverse reasoning works best under certainconditionsconditions–– Multiple goal variablesMultiple goal variables–– Many rulesMany rules–– All or most rules do not have to be examined inAll or most rules do not have to be examined in

the process of reaching a solutionthe process of reaching a solution

13-38

Development EngineDevelopment Engine

Programming languagesProgramming languages–– LispLisp–– PrologProlog

Expert system shellsExpert system shells–– Ready made processor that can be tailored to aReady made processor that can be tailored to a

particular problem domainparticular problem domain

Case-based reasoning (CBR)Case-based reasoning (CBR)Decision treeDecision tree

13-39

Expert System AdvantagesExpert System Advantages

For managersFor managers–– Consider more alternativesConsider more alternatives–– Apply high level of logicApply high level of logic–– Have more time to evaluate decision rulesHave more time to evaluate decision rules–– Consistent logicConsistent logic

For the firmFor the firm–– Better performance from management teamBetter performance from management team–– Retain firm’s knowledge resourceRetain firm’s knowledge resource

13-40

Expert System DisadvantagesExpert System Disadvantages

Can’t handle inconsistent knowledgeCan’t handle inconsistent knowledgeCan’t apply judgment or intuitionCan’t apply judgment or intuition

13-41

Keys to Successful ESKeys to Successful ESDevelopmentDevelopment

Coordinate ES development with strategicCoordinate ES development with strategicplanningplanningClearly define problem to be solved andClearly define problem to be solved andunderstand problem domainunderstand problem domainPay particular attention to ethical and legalPay particular attention to ethical and legalfeasibility of proposed systemfeasibility of proposed systemUnderstand users’ concerns and expectationsUnderstand users’ concerns and expectationsconcerning systemconcerning systemEmploy management techniques designed to retainEmploy management techniques designed to retaindevelopersdevelopers 13-42

Neural NetworksNeural Networks

Mathematical model of the human brainMathematical model of the human brain–– Simulates the way neurons interact to processSimulates the way neurons interact to process

data and learn from experiencedata and learn from experience

Bottom-up approach to modeling humanBottom-up approach to modeling humanintuitionintuition

13-43

The Human BrainThe Human Brain

Neuron -- the information processorNeuron -- the information processor–– Input -- dendritesInput -- dendrites–– Processing -- somaProcessing -- soma–– Output -- axonOutput -- axon

Neurons are connected by the synapseNeurons are connected by the synapse

13-44

Soma(processor)

Axon

Synapse

Dendrites (input)

Axonal Paths (output)

Simple Biological NeuronsSimple Biological Neurons

13-45

Evolution of ArtificialEvolution of ArtificialNeural Systems (ANS)Neural Systems (ANS)

McCullochMcCulloch--PittsPitts mathematical neuron mathematical neuronfunction (late 1930s) was the starting pointfunction (late 1930s) was the starting pointHebb’sHebb’s learning law (early 1940s) learning law (early 1940s)NeurocomputersNeurocomputers–– MarvinMarvin Minsky’s Snark Minsky’s Snark (early 1950s) (early 1950s)–– Rosenblatt’s PerceptronRosenblatt’s Perceptron (mid 1950s) (mid 1950s)

13-46

Current MethodologyCurrent Methodology

Mathematical models don’t duplicateMathematical models don’t duplicatehuman brains, but exhibit similar abilitieshuman brains, but exhibit similar abilitiesComplex networksComplex networksRepetitious trainingRepetitious training–– ANS “learns” by exampleANS “learns” by example

13-47

y1

y2

y3

yn-1

y

w1

w2

w3

wn-1

Single Artificial NeuronSingle Artificial Neuron

13-48

The Multi-LayerThe Multi-LayerPerceptronPerceptron

Yn2

ININnn

OUTOUTnnOUTOUT11

ININ11

YY11

InputInputLayerLayer

OutputLOutputLayerayer

13-49

Knowledge-based SystemsKnowledge-based Systemsin Perspectivein Perspective

Much has been accomplished in neural netsMuch has been accomplished in neural netsand expert systemsand expert systemsMuch work remainsMuch work remainsSystems abilities to mimic humanSystems abilities to mimic humanintelligence are too limited and regarded asintelligence are too limited and regarded asprimitiveprimitive

13-50

Summary [cont.]Summary [cont.]

AIAI–– Neural networksNeural networks–– Expert systemsExpert systems

Limitations and promiseLimitations and promise

13-51

Case StudyCase 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