Shravana Kumar Musunuri, Jimmy Mathews Advisors: Dr. Joseph Picone Dr. David Gao Powertrain Design Tools Project The GENERIC MODELING ENVIRONMENT (GME) INTELLIGENT POWERTRAIN DESIGN

Page 1 of 24 Intelligent Powertrain Design Outline Overview of Computer-Based Systems (CBS) Introduction to Model Integrated Computing (MIC) Principles of Domain Specific Modeling Environments (DSME) Configurable Domain Specific Modeling Environments (CDSME) Introduction to Generic Modeling Environment (GME) GME Concepts

Page 2 of 24 Intelligent Powertrain Design Computer-Based Systems Necessity o Complex systems o Efficient and Faster o Capability of Modeling and Simulation Challenges o Tight Integration of Information Processing and physical systems (prime importance) o Expensive to develop and maintain o Agility of applications

Page 3 of 24 Intelligent Powertrain Design Model-Based Systems Models: Mathematical abstractions of the behavior of physical artifacts Modeling: Usage of abstractions (models) as programming elements Model-Based Systems: Systems employing concepts of modeling o Provides basis for solving complex systems o Based on separation of problem-solving algorithm from the model

Page 4 of 24 Intelligent Powertrain Design Model-Integrated Computing Integration of Design tools and the executable system Approach for rapid design and implementation of systems where the software, the environment and the integration constraints are modeled Provides framework for developing domain artifacts for computer-based systems Relies heavily on the use of domain-specific languages to describe the final system implementation

Page 5 of 24 Intelligent Powertrain Design Model Integrated Computing (contd..) Model Integrated Program Synthesis (MIPS): One approach of model integrated computing o Operates according to a domain-specific set of requirements and describes how any system in the domain can be modeled o Specifies types of entities, relationships and requirements for domain modeling MultiGraph Architecture (MGA): Tool set for creating MIPS environments o Provides means for evolving domain-specific applications by model modifications and re-synthesis of applications.

Page 6 of 24 Intelligent Powertrain Design Domain Specific Modeling Domain: A family of related systems; e.g. engineering domain uses numerical analysis, matrix transforms, etc. Domain-specific modeling (DSM) is a technology that focuses on higher levels of abstraction at the problem space and avoids low-level details at the solutions space. Allows the user to manipulate graphical models of the problem in hand Useful in automating different kinds of applications in which the environment is dynamic and tightly integrated with the physical environment including: –embedded systems –automotive manufacturing

Page 7 of 24 Intelligent Powertrain Design Domain Specific Modeling Environment DSME: A domain-specific environment that uses models to create systems Only things related to a particular domain are available to the domain user Examples: o Matlab, Spice, Microsoft Office, AutoCAD Modeling Paradigm: Defines the family of models that can be created using the resultant MIPS environment o Provides the semantic, syntactic, presentation information regarding the domain, which are used in the construction of models.

Page 8 of 24 Intelligent Powertrain Design Domain Specific Modeling Environment (contd..) Metamodeling: Modeling of a modeling environment Metamodels: o Models of a particular modeling environment o Provides formal semantics for Domain Specific Modeling Languages Model Interpreters: o Performs the translation between domain models and applications o Traverses the model database, analyzes the models and creates the executable systems

Page 9 of 24 Intelligent Powertrain Design Domain Specific Modeling Environment (contd..) General Meta-Meta-Model Domain Meta-Model Domain Models Application Interpreter 1Interpreter 2 Application Specify Construct Generate Specific Instance Domain Specific Modeling [1]

Page 10 of 24 Intelligent Powertrain Design Domain Specific Modeling Environment (contd..) Another view Metamodeling tools are used to design a DSME. This customized environment is then used to develop the models of the system [4].

Page 11 of 24 Intelligent Powertrain Design Model-Integrated Computing-based development [2]

Page 12 of 24 Intelligent Powertrain Design Overview of MIPS Overview of MIPS [4]

Page 13 of 24 Intelligent Powertrain Design Configurable Domain Specific Modeling Environment (CDSME) Why CDSME ? o Creating a DSME for each domain is expensive and time consuming. o To include various aspects for customization of an application Example of CDSME: Generic Modeling Environment (GME), developed by Institute for Software Integrated Systems (ISIS), Vanderbilt university.

Page 14 of 24 Intelligent Powertrain Design Generic Modeling Environment The Generic Modeling Environment is a configurable tool kit for creating domain-specific modeling, model analysis and program synthesis environments.  Configuration though UML and OCL based meta- models  Extensible architecture through MS COM and.net  Multiple standard backend support (ODBC,XML)  Multiple language support: C++, VB, Java, C#, Python GME is based on the same Modeling Integrated Computing concepts like Modeling paradigm, Metamodels, MIPS.

Page 15 of 24 Intelligent Powertrain Design MIPS and GME link [7] Model Interpretation Application Domain App. 1 App. 2 App. 3 Application Evolution Environment Evolution Meta-Level Translation Metaprogramming Interface Formal Specifications Model Interpreters Models DSDE Model Builder GME

Page 16 of 24 Intelligent Powertrain Design GME 4.0 Main Editing Window [6] Title bar: Indicates the currently loaded project Menu bar: Commands for certain operations on the model Tool bar: Icon button shortcuts for several editing functions Mode bar: Buttons for selecting and editing modes Editing area: Area containing model editing windows Part browser: Shows the parts that can be inserted in the current aspect of the current model Attribute browser: Shows the attributes and preferences of an object Status bar: Shows status, error messages, current edit mode,paradigm name, zoom factor and current time Model browser: Shows either aggregation hierarchy of the project, type inheritance of model, or overview of the current modeling paradigm

Page 17 of 24 Intelligent Powertrain Design GME Concepts Models: An abstract object that represents something o Has state, identity and behavior The purpose of GME is to create and manipulate these models. E.g. A dataflow block is the model for an operator in Signal Processing domain. A model can consist of various parts like atoms, other models, references sets and connections. Default Icon

Page 18 of 24 Intelligent Powertrain Design Models containing other models as parts are called compound models. Models that cannot contain any other models as parts are called primitive models. Atoms: Simple modeling objects that do not have internal structure, but they can have attributes E.g. An output dataport on a dataflow block in Signal Processing domain. GME Concepts (contd..)

Page 19 of 24 Intelligent Powertrain Design Model Hierarchy: Models can contain other models as parts, same or different kind as the parent model. This results in model hierarchy. E.g. Hierarchical dataflow diagrams in Signal Processing domain. GME Concepts (contd..)

Page 20 of 24 Intelligent Powertrain Design GME Concepts (contd..) References: Objects that refer to other modeling objects o A reference can point to a model, an atom, a model embedded in another model or even another reference part. o Null references is possible Connections: A line that connects two parts of a model o Has two attributes, appearance and directionality o When a line is drawn, GME checks whether the connection is legal or not by determining if the two types of objects are allowed to be connected together. E.g. Connections between dataflow blocks in Signal Processing paradigm

Page 21 of 24 Intelligent Powertrain Design Links: A port through which the model is connected to another part within the parent model Aspects: Defines the kinds of parts that are visible in that aspect. The existence or visibility of a part within a particular aspect is determined by the modeling paradigm. E.g. Signal flow and ‘states’ aspects for a Signal Processing paradigm Attributes: Property of an object described textually. Objects can have multiple attributes. The modeling paradigm defines what attributes, range of values are to be present for the particular objects. E.g. Datatype of parameters in a Signal Processing paradigm GME Concepts (contd..)

Page 22 of 24 Intelligent Powertrain Design GME Concepts (contd..) Attribute box associated with a parameter atom called pi. Sets: Represent different states of a ‘dynamic system’. Is composed of almost the same parts either in ‘visible’ or ‘missing’ mode depending on the state of the system. When a particular set is activated, only objects belonging to that set are ‘visible’, the others being ‘dimmed out’.

Page 23 of 24 Intelligent Powertrain Design References 1.Hernandez.F,Bangalore.P,Gray.J,Reily.K, “ A Graphical Modeling Environment for the Generation of Workflows for the Globus Toolkit”, Workshop on Component Models and Systems for Grid Applications, June-July Nordstrom.G, Karsai.G,et.al, “Model Integrated Computing-based Software Design and Evolution”, Conference on Life Cycle Software Engineering Technology for Modern Avionics, Missiles, and Smart Weapon Systems,, Huntsville, Alabama, August Ledeczi A., “ Model Construction for Model-Integrated Computing”, 13th International Conference on Systems Engineering, Las Vegas, NV, August, Sprinkle. J, “ Model-integrated computing”, IEEE Potentials, Volume: 23, Issue: 1, February- March Sztipanovits. J, Karsai. G, “Model-integrated computing”, Computer,Volume: 30, Issue: 4, April GME 4 User’s Manual, Institute for Software Integrated Systems, Vanderbilt University. 7.Akos Ledeczi, “ The Generic Modeling Environment”, Institute for Software Integrated Systems, Vanderbilt University.

Page 24 of 24 Intelligent Powertrain Design Questions