Decision Support Systems - Chapter 4: Modeling and Analysis

Modeling languages Lingo, AMPL, GAMS (General Algebraic Modeling Systems) Relational model base management system Virtual file Virtual relation Object-oriented model base management system Logical independence Database and MIS design model systems Data diagram, ERD diagrams managed by CASE tools

ppt39 trang | Chia sẻ: dntpro1256 | Lượt xem: 646 | Lượt tải: 0download
Bạn đang xem trước 20 trang tài liệu Decision Support Systems - Chapter 4: Modeling and Analysis, để xem tài liệu hoàn chỉnh bạn click vào nút DOWNLOAD ở trên
Chapter 4 Modeling and AnalysisDecision Support Systems 1© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangOutline1. Modeling for DSS2. Static and Dynamic models3. Treating certainty, uncertainty4. Influence diagrams5. Modeling with spreadsheets6.Decision Tables and Decision trees7.MSS mathematical models8. Search approaches9.Simulation10. Model base management system2© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang1.Modeling for DSSModeling is key element in DSSMany classes of modelsSimulation is an exampleSpecialized techniques for each modelAllows for rapid examination of alternative solutionsMultiple models often included in a DSSTrend toward model transparencyMultidimensional modeling exhibits as spreadsheet3© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangSome major modeling issuesIdentification of the problem and environment analysisVariable identificationDecision variablesUncontrollable variablesResult variables, etc.Note: using influence diagrams and cognitive maps to identify variables and relationships.Forecasting: DSS is designed to determine what will be.Time series forecastingThere exist several forecasting packages.Multiple models4© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangDSS ModelsAlgorithm-based modelsStatistic-based modelsLinear programming modelsGraphical modelsQuantitative modelsQualitative modelsSimulation models5© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang6© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang2. Static and Dynamic ModelsStatic ModelsSingle snapshot of situationSingle intervalTime can be rolled forward, a snapshot at a timeUsually repeatableSteady stateOptimal operating parametersContinuousUnvaryingPrimary tool for process design7© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangDynamic ModelsRepresent changing situationsTime dependentVarying conditionsGenerate and use trends and patterns over timeOccurrence may not repeat8© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang3.Treating certainty, uncertainty and riskDecision making under certaintyAssume complete knowledgeAll potential outcomes knownEasy to developResolution determined easilyCan be very complex9© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangDecision-Making under uncertaintyUncertaintySeveral outcomes for each decisionProbability of occurrence of each outcome unknownInsufficient informationAssess risk and willingness to take itPessimistic/optimistic approaches10© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangDecision-Making under riskProbabilistic Decision-MakingDecision under riskProbability of each of several possible outcomes occurringRisk analysisCalculate value of each alternativeSelect best expected value11© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang4. Influence DiagramsGraphical representation of modelProvides relationship frameworkExamines dependencies of variablesAny level of detailsShows impact of changeShows what-if analysis12© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangInfluence DiagramsDecision Intermediate or uncontrollableVariables:Result or outcome (intermediate or final)CertaintyUncertaintyArrows indicate type of relationship and direction of influenceAmount in CDsInterest earnedPriceSales13© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangInfluence DiagramsRandom (risk)Place tilde above variable’s name ~ DemandSalesPreference(double line arrow)Graduate UniversitySleep all daySki all dayGet jobArrows can be one-way or bidirectional, based upon the direction of influence14© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangExampleProfit = income – expensesIncome = units sold  unit priceUnits sold = 0.5  amount used in advertisementExpenses = unit cost  unit sold + fixed cost 15© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang16© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang5.Modeling with SpreadsheetsFlexible and easy to useEnd-user modeling toolAllows linear programming and regression analysis (as add-ins to the software package)Features what-if analysis, data management, macrosSeamless and transparentIncorporates both static and dynamic modelsExcel and Lotus 1-2-3 are two popular spreadsheet software package.17© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang18© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang6.Decision TablesMultiple criteria decision analysisFeatures include:Decision variables (alternatives)Uncontrollable variablesResult variablesApplies principles of certainty, uncertainty, and risk19© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangTable 2 Investment Problem Decision Table Model State of Nature (uncontrollable variables) ---------------------------------------------------------------Alternative Solid growth (%) Stagnation (%) Inflation(%)-------------------------------------------------------------------------------Bonds 12.0 6.0 3.0Stocks 15.0 3.0 6.5CDs 6.5 6.5 6.520© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangDecision TreeGraphical representation of relationshipsMultiple criteria approachDemonstrates complex relationshipsCumbersome, if many alternatives21© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang7. MSS Mathematical ModelsLink decision variables, uncontrollable variables, parameters, and result variables togetherDecision variables describe alternative choices.Uncontrollable variables are outside decision-maker’s control.Fixed factors are parameters. Intermediate outcomes produce intermediate result variables.Result variables are dependent on chosen solution and uncontrollable variables.22© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangMSS Mathematical ModelsNonquantitative modelsSymbolic relationshipQualitative relationshipResults based upon Decision selectedFactors beyond control of decision makerRelationships amongst variables23© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang24© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangMathematical ProgrammingTools for solving managerial problemsDecision-maker must allocate resources amongst competing activitiesOptimization of specific goalsLinear programmingConsists of decision variables, objective function and coefficients, uncontrollable variables (constraints), capacities, input and output coefficients25© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangExample of Linear programmingObjective function: Max Z = 45x1 + 12x2Constraints: 1x1 + 1x2  300 3x1 + 0x2  250x1 and x2 are decision variablesConstraints in the form of linear inequalities or equalities.Lingo and Lindo are two best-known software packages used for Linear and Integer Programming.26© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangMultiple GoalsSimultaneous, often conflicting goals sought by managementDetermining single measure of effectiveness is difficultHandling methods:Utility theoryGoal programmingLinear programming with goals as constraintsPoint systemSome software packages for multi-goal decision making: Analytic Hierarchy Process (AHP) and Expert Choice.27© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangSensitivity, What-if, and Goal Seeking AnalysisSensitivityAssesses impact of change in inputs or parameters on solutionsAllows for adaptability and flexibilityEliminates or reduces variablesCan be automatic or trial and errorWhat-ifAssesses solutions based on changes in variables or assumptionsGoal seekingBackwards approach, starts with goalDetermines values of inputs needed to achieve goalExample is break-even point determination28© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang8. Problem solving search approachesAnalytical techniques (algorithms) for structured problemsGeneral, step-by-step searchObtains an optimal solutionBlind searchComplete enumeration All alternatives exploredIncomplete Partial searchAchieves particular goalMay obtain optimal goal29© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangSearch ApproachesHeurisiticRepeated, step-by-step searchesRule-based, so used for specific situations“Good enough” solution, but, eventually, will obtain optimal goalExamples of heuristicsHill ClimbingTabu search Remembers and directs toward higher quality choicesSimulated annealingGenetic algorithmsRandomly examines pairs of solutions and mutationsAnt colony optimization (ACO)30© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang31© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang9.SimulationsImitation of realityAllows for experimentation and time compressionDescriptive, not normativeCan include complexities, but requires special skillsHandles unstructured problemsOptimal solution not guaranteedMethodologyProblem definitionConstruction of modelTesting and validationDesign of experimentExperimentationEvaluationImplementation32© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangSimulationsProbabilistic independent variablesDiscrete or continuous distributionsTime-dependent or time-independentVisual interactive modeling GraphicalDecision-makers interact with simulated modelmay be used with artificial intelligenceCan be objected-oriented33© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang34© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangAdvantages of SimulationThe theory of a simulation is straightforward.A great amount of time compression can be attained.Managers can use a trial-and-error approach to problem solving and can do so faster, cheaper and more accuracy with simulation.Simulation helps to gain a better understanding of the problem and the potential decisions available.Simulation can handle a wide variety of problem types.Simulation is often the only DSS method that can handle unstructured problems.35© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangDisadvantages of SimulationAn optimal solution cannot be guaranteedSimulation model construction can be a slow and costly processSolutions and inferences from a simulation are not transferable to other problems.Simulation is so easy to explain to managers that analytic methods are often overlooked.Simulation software requires special skills.36© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangQuantitative software packagesStatistical packages:SPSS, Systat, SAS, TSP.Operation research packages:ILOG, OR-Objects, Lingo, Lindo, OSL (Optimization System Library), CPLEXGPSS, SIMULA, SIMSCRIPT, SLAM.Revenue Management packagesSpreadsheet add-ins37© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang10.Model-Base Management SystemSoftware that allows model organization with transparent data processingSome desirable MBMS capabilitiesDSS user has controlFlexible in designGives feedbackGUI basedReduction of redundancyIncrease in consistencyCommunication between combined models38© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and LiangModel-Base Management SystemModeling languagesLingo, AMPL, GAMS (General Algebraic Modeling Systems)Relational model base management systemVirtual fileVirtual relationObject-oriented model base management systemLogical independenceDatabase and MIS design model systemsData diagram, ERD diagrams managed by CASE tools39© 2005 Prentice Hall, Decision Support Systems and Intelligent Systems, 7th Edition, Turban, Aronson, and Liang

Các file đính kèm theo tài liệu này:

  • pptdss_chap4_1842_1810913.ppt