Presentation on theme: "MISSOURI UNIVERSITY OF SCIENCE AND TECHNOLOGY Rolla, Missouri, U.S.A."— Presentation transcript:
1 MISSOURI UNIVERSITY OF SCIENCE AND TECHNOLOGY Rolla, Missouri, U.S.A. Systems Engineering Research Taking Systems Engineering to the Next LevelCihan H Dagli, PhDProfessor of Engineering Management and Systems Engineering Professor of Electrical and Computer Engineering Founder and Boeing Coordinator of Systems Engineering Graduate ProgramINCOSE and IIE FellowMISSOURI UNIVERSITY OF SCIENCE AND TECHNOLOGY Rolla, Missouri, U.S.A.
2 Outline Introduction Academia Needs Missouri S&T’s Approach Need for Systems Architecting and EngineeringDoD Systems Engineering Vision 2020Academia NeedsMissouri S&T’s ApproachSmart Systems ArchitectingCoursesIndustry CooperationFuture Of Systems Architecting
3 Research Collaborators Renzhong Wang (Current SysEng PhD Student, INCOSE Doctoral Award Recipient)Dr. Atmika Singh (Former SysEng PhD Student, Researcher at Clearway Holding)Dr. Jason Dauby (Former SysEng PhD Student, INCOSE Doctoral Award Recipient, Researcher at Naval Surface Warfare Center)Dr. Nil Kilicay Ergin (Former SysEng PhD Student, Faculty at Pen State University)
4 Introduction The Dynamically Changing Operating Environment We are increasingly a networked society:Trans-national mega military systemsAsymmetrical threats vs. rapid reaction forcesTrans-national enterprisesTrans-national manufacturingGlobally distributed services and productionWe are increasingly dependent on these networks.
5 Introduction Effectiveness Survivability Vulnerability Mission Success Advanced SupportabilitySupply Chain Mgt.Maintenance Mgt. AnalysisSupply Mgt. AnalysisOperational C4ISRCommunicationsDynamic SystemsSystem of SystemsVisualize ScenariosImmerse Man in LoopDecision AnalysisVoice of CustomerCustomer RequirementsExpert JudgmentLCC/TOCDesign to CostBest ValueCourtesy of Dr. Mike McCoy
6 Introduction(Adopted from An Overview of Global Earth Observation System of Systems (GEOSS), Stefan Falke, Geospatial Intelligence Operating Unit, Northrop Grumman Corporation)
7 Introduction Trans-national Manufacturing Super-Efficient , Eco-Friendly, and People FriendlyTrans-national Manufacturing
8 Need for Systems Architecting and Engineering Systems Engineering: An interdisciplinary approach and means to enable the realization of successful systems. Systems Engineering considers both the business and the technical needs of all stakeholders with the goal of providing a quality product that meets the user needs.System Architecture: The aggregation of decomposed system functions into interacting system elements whose requirements include those associated with the aggregated system functions and their interfaces requirements/definitionINCOSE (International Council of Systems Engineers)
9 Need for Systems Architecting and Engineering Cost and Schedule Performance as a Function of Systems Engineering Effort*Source: INCOSE Systems Engineering Center of Excellence SECOE INCOSE 2003; & Honour, E. “Understanding Value of Systems Engineering”, INCOSE Conference, June 20-24, 2004
10 Need for Systems Architecting and Engineering Performed by NDIA in conjunction with the SEI inSurveyed 64 projects at defense contractors to assess:Data was collected anonymously to encourage honest and accurate reporting.Results published at:ocuments/08.reports/08sr034.html*Source: Presentation of Joe Elm from Software Engineering Institute Carnegie Mellon University
11 Need for Systems Architecting and Engineering PROJECT PERFORMANCE vs. TOTAL SE CAPABILITY39%46%15%29%59%12%31%13%56%BestPerformance( x > 3.0 )Moderate( 2.5 x 3.0 )Lower( x < 2.5 )Lower Capability( x 2.5 )N = 13Moderate Capability( 2.5 < x < 3.0 )N = 17HigherCapability(x 3.0 )N = 16Gamma = 0.32p = 0.041.000.750.500.250.00*Source: Presentation of Joe Elm from Software Engineering Institute Carnegie Mellon University
12 Need for Systems Architecting and Engineering *Source: Presentation of Joe Elm from Software Engineering Institute Carnegie Mellon University
13 Need for Systems Architecting and Engineering *Source: Presentation of Joe Elm from Software Engineering Institute Carnegie Mellon University
14 Need for Systems Architecting and Engineering Architectures are fundamental to the success of the programArchitecture selection is a search process based on ambiguous information and dataArchitecture selection requires assessment methods based on ambiguous key performance parameters to identify compromise architectureArchitecting process is reduction of ambiguity hierarchically
15 DoD Systems Engineering Vision 2020 Design PrinciplesPlatform Based Engineering Using a common core platform to develop many related systems/capabilitiesTrusted System Design Developing trusted systems from untrusted components
16 DoD Systems Engineering Vision 2020 Design FrameworkModel Based Engineering Using modeling and simulation for rapid, concurrent, integrated system development and manufacturing
17 DoD Systems Engineering Vision 2020 Adaptable DoD SystemsCapability on Demand Real-time Adaptive Systems Rapidly Reconfigurable Systems Pre-planned Disposable Systems
18 Academia NeedsSystems Architecting Laboratory: Real Engineering Problems and CustomerEnvironment to demonstrate, value of systems engineering and new systems architecting approaches on real systems of various sizeClose cooperation with industry honoring propriety nature of information and dataDissemination channels for new research
19 Missouri S&T’s Approach Systems Architecting Research
20 Smart System Architecting How can we assess architectures?How can we represent architectures?How can we generate architectures?How can we reduce ambiguity hierarchically?How can we test architectures for correctness?What are the tools of architect?
21 Smart Systems Architecting C. H. Dagli, A. Singh, J. P. Dauby, R. Wang, “Smart systems architecting: computational intelligence applied to trade space exploration and system design,” Systems Research Forum ,Vol. 3, No. 2 (2009) 101–119
22 DoD Systems Engineering Vision 2020 Design FrameworkModel Based Engineering Using modeling and simulation for rapid, concurrent, integrated system development and manufacturing
23 Smart Systems Architecting What constitutes the “best” in architecture?What is the measure for comparing architectures?We can search for the “best” architecture, as long as we can define “best”Can we associate an aggregate value in evaluating functional architectures?How can we deal with the ambiguity of need requirements and performance measures in the search process?Is there a way to mathematically represent functional architectures?Can we generate architectures through a evolutionary process?Can we integrate the architect in evolutionary architecting process?C. H. Dagli, A. Singh, J. P. Dauby, R. Wang, “Smart systems architecting: computational intelligence applied to trade space exploration and system design,” Systems Research Forum ,Vol. 3, No. 2 (2009) 101–119
24 Smart Systems Architecting PERFORMANCEPERCEPTIONSSCHEDULECOSTRISKFACTSWhat is the measure for comparing architectures?
25 Smart Systems Architecting AdaptabilityAffordabilitySurvivabilityRobustnessFlexibilityReliabilityWhat is a reasonable approach to find and aggregate measure for comparing architectures?
26 Smart Systems Architecting Super-Efficient , Eco-Friendly, and People FriendlyTop level system attributes
27 Smart Systems Architecting (SSA) SSA ApproachFuzzy Assessment and Computing with wordsEvolutionary Algorithms for ArchitectureCanonical Decomposition Fuzzy Comparison (CDFC)Self Organizing Maps for Clustering Architecture FamiliesModels for Behavior ModelingC. H. Dagli, A. Singh, J. P. Dauby, R. Wang, “Smart systems architecting: computational intelligence applied to trade space exploration and system design,” Systems Research Forum ,Vol. 3, No. 2 (2009) 101–119
28 Fuzzy Assessment and Computing with Words Modern large-scale systems are comprised of many interacting subsystems and components and exhibit complex behavior.This nonlinear behavior cannot be analyzed using traditional modeling approaches.Fuzzy Cognitive Maps based methodology can be for assessing the inherent value of candidate architectures early in the design lifecycle.A. Singh and C. H. Dagli, “"Computing with words" to support multi-criteria decision making during conceptual design,” Systems Research Forum, vol. 04, no. 01, p. 85,
29 Fuzzy Assessment and Computing with Words The system and its components are represented in the form of a directional graph where the nodes represent system components and the arcs represent their interactions.This modeling approach makes use of the “computing with words” (CW) paradigm to use human experience to assign linguistic weights to the arcs based on the strength of influence between connected nodes.An overall value measure for a system can be derived by simulating the resulting graph. Such an approach will facilitate the selection of the best set of architectures or component technologies during the nascent design stages based on the value delivered to the stakeholder.A. Singh and C. H. Dagli, “"Computing with words" to support multi-criteria decision making during conceptual design,” Systems Research Forum, vol. 04, no. 01, p. 85,
30 Evolutionary Algorithms for Architecture Once architecture options have been identified using FCM and CW, evolutionary algorithms can be employed to find the right combination of technologies to utilize in a system design.Functional architecture chromosome
31 Canonical Decomposition Fuzzy Comparison (CDFC) The CDFC methodology is a new architecture assessment approach offering increased objectivity, fidelity, and defensibility in comparison to traditional approaches.The methodology consists of four elements:Extensible modeling – facilitates the exchange of data between model resolution levels.Canonical design primitives – basic representations of system-component technologies.Comparative analysis – comparison between heuristic and canonical embodiments.Fuzzy inference – a mapping from system response features to fuzzy sets describing the architecture assessment.J. P. Dauby, “Assessing system architectures: the canonical decomposition fuzzy comparative methodology,” Ph.D. dissertation, Dept. Eng. Management and Sys. Eng., Missouri University of Science and Technology, Rolla, MO, 2011.
32 Canonical Decomposition Fuzzy Comparison (CDFC) Architecture assessment for airborne wireless systems in conjunction with a potentialAcquisition Category (ACAT) ID program for the Department of the NavyJ. P. Dauby, “The canonical decomposition fuzzy comparative approach to assessing physical architectures,” INSIGHT, vol. 13, no. 3, pp , Oct
33 Self Organizing Maps for Clustering Architecture Families Architecture solution candidates are described by functional, logical, or physical properties including integration sensitivity.The set of properties for each candidate are used as the input vector to a variety of SOM algorithms.The SOM output can identify design features and group potential architectural concepts into families based on common features, sensitivities, or tendencies.This approach facilitates the development of architecture families that exhibit similar behavior as well as identify combinations of technologies that work well together.
34 Models for Behavior Modeling MotivationIntroduce dynamic model analysis into architecture modeling.Facilitate system behavior, performance, and effectiveness analysis, architecture evaluation, and functionality verification and validation.Renzhong Wang and Cihan H. Dagli, “Executable System Architecting Using Systems Modeling Language in Conjunction with Colored Petri Nets in a Model Driven Systems Development Process.” Journal of Systems Engineering, Vol. 14(3), 2011
35 Models for Behavior Modeling Requirement Analysis and SpecificationRequirements AnalysisFormal ModelSysML DiagramsExecutable modelCPNSimulationInteractive GUIArchitecture Analysis and EvaluationArchitecture refinement & reconfigurationFunctionality verificationBehavior analysisStartEndModel TransformationBehavior as modeledDesired BehaviorRefinementExternal ApplicationRenzhong Wang and Cihan H. Dagli, “Executable System Architecting Using Systems Modeling Language in Conjunction with Colored Petri Nets in a Model Driven Systems Development Process.” Journal of Systems Engineering, Vol. 14(3), 2011
36 Models for Behavior Modeling OMG (Object Management Group), Semantics of a Foundational Subset for Executable UML Models, Version 1.0 Beta 3, ptc/ , aFoundational UML Reference Implementation,Specify and demonstrate the semantics required to execute activity diagrams and associated timelines per the SysML v1.0 specificationSpecify the supporting semantics needed to integrate behavior with structure and realize these activities in blocks and parts represented by activity partitionsRenzhong Wang and Cihan H. Dagli, “Executable System Architecting Using Systems Modeling Language in Conjunction with Colored Petri Nets in a Model Driven Systems Development Process.” Journal of Systems Engineering, Vol. 14(3), 2011
37 Models for Behavior Modeling “Behavioral Formalism” refers to a formalized framework for describing behavior, such as state machines, Petri nets, data flow graphs, etc.UML/SysML, modeling language weak in executable semanticsSupplemented by Semantics of a Foundational Subset for Executable UML ModelsSoftware that implemented behavioral formalismCORE, IBM Rational Rhapsody, CPN Tools, etc.Renzhong Wang and Cihan H. Dagli, “Executable System Architecting Using Systems Modeling Language in Conjunction with Colored Petri Nets in a Model Driven Systems Development Process.” Journal of Systems Engineering, Vol. 14(3), 2011
38 Models for Behavior Modeling Combined usage of related tools.Three basic functions of a model:Specification (of a system to be built),UML and SysMLPresentation (of a system to be explained to other people, or ourselves),DoDAF productsSimulation.Petri nets, DEVS (Discrete Event Specification SystemxUML, XTUML, VM, Business Process Modeling Notation/Business Process Execution Language BPMN /BPELExtract key information from simulation to support architecture evaluation and analysis.Refine the architecture design based on analysis results.Renzhong Wang and Cihan H. Dagli, “Executable System Architecting Using Systems Modeling Language in Conjunction with Colored Petri Nets in a Model Driven Systems Development Process.” Journal of Systems Engineering, Vol. 14(3), 2011
39 DoDAF 2.0 Architecture Viewpoints and DoDAF-described Models DoD Architecture Framework Version 2.0 Volume I
40 Architecture Presentation Techniques DoD Architecture Framework Version 2.0 Volume I
41 Architecture Analytics DoD Architecture Framework Version 2.0 Volume I
42 Executable Modeling Formalisms The chosen of executable modeling language depends on the system to be modeled, the abstraction level to work on, and the system behavior of interest.Many modern distributed systems can be best specified by discrete event models becauseThe behavior of these systems is driven only by events that occur at discrete time points.Discrete-event models* represent the operation of a system as a chronological discrete sequence of events. Each event occurs at an instant in time and marks a change of state in the system.An executable architecture is a dynamic model that defines the precise event sequences, the conditions under which event is triggered and information is produced or consumed, and the proprieties of producers, consumers and other resources associated with the operation of the system.* Banks, J. Discrete-event System Simulation. Pearson Prentice Hall, Upper Saddle River, NJ
43 Colored Petri Nets (CPNs) Places carry makers, called tokens, which mark the state of a system.Transitions describe the actions of the systemWhen certain conditions hold, transitions will be fired, causing a change in the placement of tokens and thus the change of system states.Arcs tell how actions modify the state and when they occur11`”data”A binding Transition is a transition that is ready to fire. It requires there are sufficient tokens in each of the input places and the tokens match the arc expressions (Inscriptions)CPNs offer an advantage of combiningCombining a well-defined mathematical foundation, an interactive graphical representation, and the capabilities to carry out simulations and formal verifications.The same models can be used to check both the logical or functional correctness of a system and for performance analysis.CPNs are very flexible in token definition and manipulation.
44 Executable Semantics Place (w tokens) Place (w tokens) CPN Transition SystemConditionsEventEffectsInput Data/Information Control signals Resources OtherAction /Activity (a set of actions)Output Data/Information Control signals Resources OtherTime DelayPost conditionsDiscrete Event System SpecificationTransitionStateStateRelationships between CPN Artifacts, System Entities and Discrete Event System Specifications
45 Models for Behavior Modeling System 1Meta-ArchitectureSystem 4G I GNet-Centric ArchitectureRobustInteroperableAdaptableFlexibleModularSystem 2Systemn-1System 3System nDynamically Changing Meta-Architecture for Complex Systems
46 Models for Behavior Modeling For modeling the meta-architectureMulti-agent based modelingAgentsEnvironmentInteractionsFor modeling sub-system architecturesCognitive architecturesN. Kilicay-Ergin “Architecting System of Systems: Artificial Life Analysis of Financial Market Behavior”, PhD Dissertation Dept. Eng. Management and Sys. Eng., Missouri University of Science and Technology, Rolla, MO, 2007
47 Deliberative Reasoning SwarmIntelligenceReinforcementLearningGeneticAlgorithmNeuralNetworksComputationalToolboxShort-termmemoryLong-termAssociative memoryAttention filterBiasImitationMechanismModulesReactive MechanismDeliberative ReasoningMeta-managementPerceptionActionAgent 1= System 1Agent 2= System 2Agent 3= System 3Agent n= System nCognitiveLevel*AgentLevelEnvironmentSystem Level BehaviorClassifiersDynamicsSemanticsSelection CriteriaSystem-of-SystemsMeta-architectureSub-system architectures*Sloman’s H-Cogaff architecture, 2000
48 Missouri S&T’s Approach Degrees and Graduate Certificates
49 Systems Engineering Degrees MS in Systems EngineeringArchitected based on a need statement of invited Boeing RFP in 1998.Since the inception of the program on Spring 2000 semester 410 engineers have received their M.S. degrees.Ten courses – six core and four engineering specialization- are required for the degree.PhD in Systems EngineeringOne graduate from Boeing Seattle out of four graduates since 2006Fifteen students currently in the program
50 Systems Engineering MS Degree CurriculumCore CoursesSystemsSystems ArchitectureSysEng 469 – Systems ArchitectingSystems Engineering and AnalysisSysEng 368 – SystemsEngr. and Analysis ISystems Engineering – Information Based DesignSysEng 468 – SystemsEngr. and Analysis IIComplex Systems ManagementEconomic Decision AnalysisSysEng 413 Economic Analysis for Systems EngineeringSystems Engineering Mgt.SysEng 412 Complex Engineering Systems Program Mgt.Organizational Behavior and ManagementSysEng 411 Systems Engineering Capstone
51 Systems Engineering Graduate Certificates Network Centric Graduate CertificateComputational Intelligence Graduate CertificateModel Based Systems Engineering Graduate Certificate (In Approval Process )Software Architecting and Engineering Graduate Certificate
52 Systems Engineering Graduate Certificate SysEng 368 Systems Engineering and Analysis I SysEng 468 Systems Engineering and Analysis II SysEng 413 Economic Analysis for Systems Engineering SysEng 469 Systems Architecting Students completing these four courses with a minimum grade of B in each course are admitted to the M.S. degree program in Systems Engineering without taking the GRE.
53 Network Centric Systems Graduate Certificate Core Courses:SysEng/CpE 419 Network-Centric Systems Architecting and EngineeringCpE/SysEng 449 Network-Centric Systems Reliability and SecurityCommunications Engineering Elective Courses (select two):CpE 317 Fault Tolerant Digital SystemsCpE 319 Digital Network DesignCpE 349 Trustworthy, Survivable Computer NetworksCpE/SysEng 348 Wireless NetworksCpE /SysEng 443 Wireless Adhoc and Sensor Networks CpE 448 High Speed NetworksCS 483 Computer SecurityCS 486 Mobile and Sensor Data Management
54 Computational Intelligence Graduate Certificate Core Courses:CpE 358/EE367/SysEng 367 Computational Intelligence and select one of the following: CS 347 Introduction to Artificial IntelligenceCS 348 Evolutionary ComputingSysEng 378/CS 378/EE 368 Introduction to Neural Networks and Applications Elective Courses (Select two courses not taken as a core course):EE/CpE/Sys Eng 301 Evolvable HardwareCS 447 Advanced Topics in Artificial IntelligenceCS 448 Advanced Evolutionary ComputingSysEng/CpE/EE 458 Adaptive Critic DesignsCS/SysEng/CpE 404 Data Mining and Knowledge DiscoveryEE 337 Neural Networks for ControlSysEng 378/CS 378/EE 368 Introduction to Neural networks and ApplicationsCpE/SysEng/EE 457 Markov Decision ProcessesSysEng 478 Advanced Neural Networks
55 Model Based Systems Engineering Graduate Certificate SysEng 433 Distributed Systems ModelingSysEng 435 Model Based Systems EngineeringSysEng 479 Smart Engineering Systems DesignEmgt 374 Engineering Design Optimization
56 Software Architecting and Engineering Graduate Certificate CS 308 Object Oriented Analysis and DesignCs 309 Software Requirements EngineeringSysEng 435 Model Based Systems EngineeringSysEng 470 Software Intensive Systems Architecting
57 Research CooperationDARPA Manufacturing Experimentation and Outreach (MENTOR) Program supplier to Boeing Research and Technology- Awarded, Duration: One yearDepartment of Defense Systems Engineering Research Center- University Affiliated Research Center SERC-UARC at Stevens Institute of Technology Project “Agile Systems Engineering: Experiential and Active Learning Approach”, Duration: 05/15/2010 to 7/31/2011Department of Defense University Affiliated Research Center for Systems Engineering Research Joint Proposal with Steven’s Institute of Technology, University of Southern California and other participating universities. October 2008 – October 2013The Boeing Company, Systems Engineering MS Degree Program for Italian Engineers: Under Industrial Return Project Italian 767 Tanker Transport, BOEING Industrial Participation Program Duration: 2006 – 2009
58 Future Research NeedsAs an integrated global society, we depend on complex, distributed engineering systems that can adapt to the dynamically changing needs of society.These systems are seen in health care, infrastructure, transportation, energy, defense, security, environmental, manufacturing, communications and supply chain systems, among others.Adaptability within these systems is critical. We need to push the boundaries of research in Complex Adaptive Systems and respond to the continuous global change in systems needs.
59 Recent PublicationsRenzhong Wang and Cihan H. Dagli, “Executable System Architecting Using Systems Modeling Language in Conjunction with Colored Petri Nets in a Model Driven Systems Development Process.” Journal of Systems Engineering, Article first published online: 4 March 2011Dauby, J. P. Dagli, C. H. , "The Canonical Decomposition Fuzzy Comparative Methodology for Assessing Architectures," Systems Journal, IEEE , vol.5, no.2, pp , June 2011Aaron A. Tucker, Gregory T. Hutto and Cihan H. Dagli “ Application of Design of Experiments to Flight Test: A Case Study” Journal of Aircraft Vol. 47, No.2,March-April 2010Atmika Singh and Cihan H Dagli ““ Computing with words” to Support Multi-Criteria Decision-Making During Conceptual Design” Systems Research Forum Vol. 4, No. 1 (2010)
60 Recent PublicationsC.H. Dagli, Atmika Singh, Jason P. Dauby and Renzhong Wang “ Smart Systems Architecting: Computational Intelligence Applied to Trade Space Exploration and System Design”, Systems Research Forum Vol. 3, No. 2 (2009) 101–119.A.A. Tucker and C.H. Dagli, "Design of Experiments as a Means of Lean Value Delivery to the Flight Test Enterprise”, Journal of Systems Engineering, volume 12, Number 3, PpM. Rao, S. Ramakrishnan, and C. Dagli, “Modeling and simulation of net centric system of systems using systems modeling language and colored Petri-nets: A demonstration using the global earth observation system of systems,” Systems Engineering, vol. 11, 2008, pp
61 Concluding RemarksMost biological systems do not forecast or schedule They respond to their environment — quickly, robustly, and adaptivelyAs engineers, let us don’t try and control the system. Design the system so that it controls and adapts itself to the environment created by dynamically changing needs