Abstraction made Concrete Hans Vangheluwe (with Pieter Mosterman, Bentley Oakes, Ahsan Qamar, Ken Vanherpen and Joachim Denil http://msdl.cs.mcgill.ca/conferences/CAMPaM/2015/ discussions) MPM4CPS Training School Tallinn, Estonia 21 March 2016
at the most appropriate level(s) of abstraction at the most appropriate level(s) of abstraction using the most appropriate formalism(s) using the most appropriate formalism(s) explicitly modelling processes explicitly modelling processes Enabler: (domain-specific) modelling language engineering, Enabler: (domain-specific) modelling language engineering, including model transformation including model transformation Pieter J. Mosterman and Hans Vangheluwe. Computer Automated Multi-Paradigm Modeling: An Introduction. Simulation: T ransactions of the Society for Modeling and Simulation International , 80(9):433- 450, September 2004. Special Issue: Grand Challenges for Modeling and Simulation.
http://dsm-tp.org
Deployment onto AUTOSAR Automating transformations , modelling and simulation -based design-space exploration Co-simulation (MIL, SIL, HIL)
28 different modelling formalisms 50 transformations FTG+PM: An Integrated Framework for Investigating Model T ransformation Chains, Levi Lûcio, Sadaf Mustafjz, Joachim Denil, Hans Vangheluwe, Maris Jukss. Proceedings of the System Design Languages Forum (SDL) 2013, Montreal, Quebec. Lecture Notes in Computer Science (LNCS), Volume 7916, pp 182-202, 2013.
FTG+PM (model mgmt. … consistency )
Formalism Transformatjon Transformatjon Executjon Transformatjon Defjnitjon (1 rule)
Classifying The Real World What is an Ontology? 8
“Linguistjc” Reasoning Ontological Reasoning in MBSE (e.g., simulatjon) Reaction time < 1 ms True Linguistic World Ontological World Safe? Reaction? Real World (RW) Linguistically conforms to Ontologically conforms to Represents Transforms Checks satisfaction Conforms to Requires Based on: [4] B. Barroca, T. Kủhne, and H. Vangheluwe. Integratjng language and ontology engineering. In MPM ’14, volume 1237 of CEUR, pages 77–86, September 2014. 9
Consistency Linguistjc – Ontological Ontological reasoning in MBSE Based on: [4] B. Barroca, T. Kủhne, and H. Vangheluwe. Integratjng language and ontology engineering. In MPM ’14, volume 1237 of CEUR, pages 77–86, September 2014. 10
Causes of Problems Ontological Reasoning in MBSE Three fundamental relationships in design processes 5 : • Multi-Semantics (MS) • Multi-Abstraction (MA) • Multi-View (MV) [5] K. Vanherpen et al. Ontological Reasoning for Consistency in the Design of Cyber-Physical Systems. Submitued to Cyber-Physical Productjon Systems. 11
Ontological Reasoning in MBSE Multj-View Example Multj-View (MV) – example 12
Ontological Reasoning in MBSE Multj-View Multj-View (MV) SD I LTM I LTM II SD II [[.]] [[.]] performance value I (PV I ) model I model II performance value II (pv II ) [[.]] [[.]] Linguistic World Ontological World Prop I = f(pv I ) Prop II = f(pv II ) Real World (RW) Holds Represents Linguistically conforms to Transforms Checks satisfaction Conforms to 13
Ontological Reasoning in MBSE Multj-Semantjcs Multj-Semantjcs (MS) SD I LTM SD II [[.]] [[.]] performance value I (pv I ) model performance value II (pv II ) [[.]] [[.]] Linguistic World Ontological World Prop I = f(pv I ) Prop II = f(pv II ) Real World (RW) Holds Represents Linguistically conforms to Transforms Checks satisfaction Conforms to 14
Ontological Reasoning in MBSE Multj-Abstractjon Multj-Abstractjon (MA) SD I LTM SD II [[.]] [[.]] A(model) performance value A(M) (pv A(M) ) [[.]] model performance value I (pv M ) [[.]] Linguistic World Ontological World Prop M = f(pv M ) Prop A(M) = f(pv A(M) ) Real World (RW) Holds Represents Linguistically conforms to Transforms Checks satisfaction Conforms to 15
Linking Linguistjc and Ontological Power Window – Ontology This results in an ontology which allows us to reason at the same level about: – Multi-Semantics – Multi-Abstraction – Multi-View 16
Conceptual world: M1 is an abstraction of M2 with respect to P if for all p in P: M1 |= p -> M2 |= p.
● Measure of correspondence - How the information from experiments in the produced system relate to the information from experiments in the original system ● Precision - The level of detail in the correspondance ● Approximation - Planned or unplanned reduced precision ● Refjnement - The adding of new information in the produced system ● Discrepancies - Features of the produced system where there is insuffjcient precision ● Fidelity - A lack of discrepancies for all information of interest
Recommend
More recommend
