Building a General Resource ModelThe COCOMO Family of ModelsMODELING AND MEASURING RESOURCESCOCOMO: Static Adjusted BaselineMODELING AND MEASURING RESOURCESCOCOMOMODELING AND MEASURING RESOURCES Levels of the COCOMO ModelMODELING AND MEASURING RESOURCES Modes of the COCOMO ModelMODELING AND MEASURING RESOURCES Formulas the Basic COCOMO ModelMODELING AND MEASURING RESOURCES Formulas the Basic COCOMO ModelMODELING AND MEASURING RESOURCES The Intermediate COCOMO ModelINTERMEDIATE COCOMO MODELHow to Use the Cost DriversMODELING AND MEASURING RESOURCESFormulas of the Intermediate COCOMO ModelMODELING AND MEASURING RESOURCESFormulas of the Intermediate COCOMO ModelMODELING AND MEASURING RESOURCESSource Code Size Used in the COCOMO ModelSOFTWARE DEVELOPMENT EFFORT MULTIPLIERSSOFTWARE DEVELOPMENT EFFORT MULTIPLIERSMODELING AND MEASURING RESOURCESOther Parameters Used in the COCOMO ModelMODELING AND MEASURING RESOURCESCOCOMO IIExcerpt from:Building Parametric ModelsOutlineRange of SE Parametric ModelsParametric Model FormsGoals: Model Success CriteriaUSC-CSE Modeling MethodologyStep 1: Determine Model NeedsMajor Cost/Schedule Decision Situations(Helped by COCOMO II)Step 2: Analyze Existing LiteratureNonlinear Reuse EffectsReuse Cost Increment forSoftware Understanding (SU)Step 3: Perform Behavioral AnalysisUSC-CSE Modeling MethodologyStep 4: Relative Significance: COSYSMOStep 4 COCOMO II Result: New Scaling Exponent ApproachStep 4: Define Relations, Data, Rating ScalesStep 5: Initial Delphi AssessmentSize Drivers vs. Effort MultipliersStep 6: Gather, Analyze Project DataInitial Data Analysis May Require Model RevisionCOCOTS Effort Distribution: 20 ProjectsRevised COCOTS ModelNew Glue Code Sub-model ResultsStep 7: Bayesian CalibrationResults of Bayesian Update:Using Prior and Sampling InformationCOCOMO II. 2000 Productivity RangesUSC-CSE Modeling MethodologyCOCOMO II Experience Factory: ICOCOMO II Experience Factory: IICOCOMO II Experience Factory: IIICOCOMO II Experience Factory: IVStep 8 Example: Software Understanding Increment Too LargeConclusionsGoals: Model Success CriteriaHOW TO USE THE MODELSReferencesBuilding a General Resource ModelThe COCOMO Family of ModelsMODELING AND MEASURING RESOURCESCOCOMO: Static Adjusted Baseline Static single variable effort equation acts as a baseline equation, e.g., effort = A * sizebThis provides a basic estimate of effortThe initial estimate is adjusted by a set of multipliers that attempts to incorporate the effect of important product and process attributesE.g, if the initial estimate is E = 100 staff months and the complexity of the job is rated higher than normal, a multiplier 1.1 is associated with it, yielding an adjusted estimate of 110 staff monthsMODELING AND MEASURING RESOURCESCOCOMO Effort = A * sizebA + B determined by regression analysis (E = 2.3 L1.27)Product:Fault freedom, data base size, product complexity, adaptation from existing softwareComputer:Execution time constraints, machine storage constraints, virtualmachine volatility, computer response timePersonnel:Analyst capability, applications experience, programmer capability, virtual machine experience, programming language experienceProject:Modern programming practices, use of software tools, required development scheduleBoehm/TRWMODELING AND MEASURING RESOURCESLevels of the COCOMO ModelThe COCOMO model levels are:- Basic, which is used for quick, early approximateestimates of software cost and schedule, but its accuracy is limited due to not using detailed data.- Intermediate, which is used for better estimates of cost and schedule, because it considers software project environment factors in terms of their aggregate impact on the project parameters.- Detailed, which is used for even better estimates, because it accounts for the influence of the software project environment factors on individual project phases.We will concentrate our discussions on the basic and intermediate levels.MODELING AND MEASURING RESOURCESModes of the COCOMO ModelThe COCOMO model modes are:- Organic mode, which is appropriate for small, stable projects- Embedded mode, which is appropriate for large projects with tight constraints, that require some degree of innovation and have a complex software interface- Semi-detached mode, which is appropriate for projects that fall in between the above two categoriesMODELING AND MEASURING RESOURCESFormulas the Basic COCOMO ModelThe required effort (E) to develop the software system as a function of source size (S) (where E is expressed in Person-Months and S is expressed in KLOC):BASIC COCOMO MODEL________________________________________________MODE EFFORT________________________________________________Organic E = 2.4 * (S1.05)_________________________________________________Semi-detached E = 3.0 * (S1.12)_________________________________________________Embedded E = 3.6 * (S1.20)_________________________________________________MODELING AND MEASURING RESOURCESFormulas the Basic COCOMO ModelThe project duration (TDEV) as a function of effort (E) (where TDEV is expressed in calendar months, and E in Person-Months):BASIC COCOMO MODEL_____________________________________________________MODE SCHEDULE______________________________________________________Organic TDEV = 2.5* (E0.38)______________________________________________________Semi-detached TDEV = 2.5 * (E0.35)______________________________________________________Embedded TDEV = 2.5 * (E0.32)_______________________________________________________MODELING AND MEASURING RESOURCESThe Intermediate COCOMO ModelThe intermediate COCOMO is an extension of the basic COCOMO model, which used only one predictor variable, the KLOC variableThe intermediate COCOMO uses 15 more predictor variables called “cost drivers.” The manager assigns a value to each cost driver from the range:Very lowLowNominalHighVery highExtra highFor each of the above values, a numerical value corresponds, which varies with the cost driversINTERMEDIATE COCOMO MODELHow to Use the Cost DriversThe manager assigns a value to each cost driver according to thecharacteristics of the specific software projectThe numerical values that correspond to the manager assigned values for the 15 cost drivers are