Tuesday 14 June 2022

12:15 -- 15:00 presentations/discussions

• 12:45 -- 13:45 Jérôme Hugues (SIE/CMU, USA)
  Title: Multi-Paradigm Modeling and simulation, an architectural perspective.
  Presentation [pdf]
  Abstract: Complex systems require multiple forms of modeling artefacts to describe various aspects: static architecture, behavior, performance analyses, safety, etc. The SAE AADL standard has been defined to represent the architecture of complex cyber-physical systems. In this talk, I will review some past projects I conducted at ISAE and SEI that illustrate the following thesis:
  • A generic agnostic architecture supports the integration of heterogeneous models. With examples from the ESA TASTE toolchains, we show how to address interoperability issues when combining models for data exchange and execution
  • There is a continuum between model checking, model simulation and execution on platform. Yet there are few works on unifying these three views.
• 13:45 -- 14:30 Dominique Blouin (Telecom Paris, France)
  Title: Hierarchical Megamodels for Model Management in Architecture-Centric Virtual Integration Development.
  Presentation [pdf]
  Abstract: During this talk, I will present the ongoing ACMoM (Architecture-Centric Model Management) project and its model management infrastructure applied to the development of embedded system with the AADL language.
• 14:30 -- 15:00 Rakshit Mittal (BITS Goa, India)
  Title: OSATE-DIM Solves the Instance Model View-Update Problem in AADL
  Presentation [pdf]
  Abstract:
  The Architecture Analysis and Design Language (AADL) is a rich language for modeling embedded systems through several constructs such as component extension and refinement to promote modularity of component declarations. To ease processing AADL models, OSATE, the reference tool for AADL, defines another model (namely `instance' model) computed from a base `declarative' model/s. An instance model is a simple object tree where all information from the declarative model is flattened so that tools can easily use this information to analyze the system. However for modifications, the tools have to make changes in the complex declarative model since there is no automated backward transformation (de-instantiation) from instance to declarative models. Since the instance model is a `view' of the declarative model, this is a view-update problem. In this talk, I will present the OSATE Declarative-Instance Mapping Tool, ( OSATE-DIM), an Eclipse plugin for de-instantiation of AADL models implementing a solution of this view-update problem. I will also present tests for OSATE-DIM with a benchmark of existing AADL model processing tools and verify the correctness of our de-instantiation transformations.

15:00 -- 15:30 coffee break (Jura machine at work :)

15:30 -- 17:30 presentations/discussions

• 15:30 -- 16:00 Arkadiusz Rys (UAntwerpen) and Randy Paredis (UAntwerpen)
  Title: Extending the FTG+PM (or at least explaining its "tentacles")
  Presentation [pdf]
  Abstract: We explain the use of the FTG+PM in models based systems engineering by going into detail on the use case of an application specific line following robot. The concept of traceability is applied in regard to different aspects of the use case and its benefits are made clear. To do this we also explain the addition of the process trace and a storage framework to the FTG+PM formalism. The process trace allows tracking of activities and their attached artifact versions and is the result of following the activities described in the process model.
• 16:00 -- 16:30 Randy Paredis (UAntwerpen)
  Title: Specifying and simulating hybrid modelling languages: the combination of (embedding in) ODE/CT-CBD and TFSA by mapping them onto DEVS
  Presentation [pdf]
  Abstract: Multi-Paradigm Modeling (MPM) advocates to explicitly model every part and aspect of a system, at the most appropriate level(s) of abstraction, using most appropriate formalism(s).  We show, starting from a representative Personalized Rapid Transportation rail car example, how MPM naturally leads to the need to combine formalisms. To give these formalisms a precise semantics and to make them executable, we choose to map them all onto behaviourally equivalent (modulo some level of approximation in the case of continuous formalisms) Discrete EVent system Specification (DEVS) models.  Our focus and main contribution is the principled combination TFSA>(CBD+StEL) of Timed Finite State Automata (TFSA) and Causal Block Diagrams (CBDs) using  a State Event Location "glue" formalism StEL, and their mapping onto DEVS. The result of our principled workflow, explicitly modeled in a Formalism Transformation Graph + Process Model, is an accurate and efficient simulator. This is demonstrated on the rail car case.
• 16:30 -- 17:00 Joeri Exelmans (UAntwerpen)
  Title: Optimistic Versioning for Blended Modeling (aka "Onion Versioning")
  Presentation [pdf]
  Abstract: Versioning systems are a fundamental piece of technology in any type of collaborative work. In MDE, textual, snapshot-based versioning systems (like Git) are ill-suited because they only record serializations of (often visual) models, introducing a lot of accidental complexity. Existing model versioning research aims to improve this by reasoning about model history (detecting changes, merging, resolving conflicts, etc.) at the level of abstract syntax, which is closer to the meaning/intend of a model. However, they support only trivial mappings between concrete and abstract syntax. To solve this, we propose a versioning data structure that not only records the history of concrete and abstract syntax, but also the history of the correspondences between them. This way, we can keep separate the orthogonal complexities of (1) optimistic versioning (branching, merging, conflict detection/resolution) and (2) non-trivial synchronizations between concrete and abstract syntax. We show by means of a minimal representative running example, that our approach is especially well suited to support Blended Modeling, a recent area of research, where the same model can be edited through multiple concrete syntaxes, possibly concurrently.
• 17:00 -- 17:30 Joeri Exelmans (UAntwerpen) and Moussa Amrani (UNamur)
  Title: Exploring the Dimensions of Language Composition
  Presentation [pdf]
  Abstract: In language engineering, language composition is the act of combining (parts of) existing, simpler modeling languages into new modeling languages, sometimes called hybrid languages. In industry today, hybrid languages and their modeling+simulation tools are developed in an ad-hoc manner. This makes it costly to adapt/evolve existing languages, or to create new languages. Ultimately, a scientific framework is needed to formally describe hybrid languages (and their tools) as combinations of simpler languages (and their "tool fragments"). As a first step, we explore the different dimensions of language composition, illustrated with examples. We see that some combinations are trivial, and others not so trivial.

18:15 -- 19:00 travel to city centre by tram 7 (or bike)
18:00 -- 19:00 the Kulminator closes at 19:00, so beer at Paters Vaetje
20:00 -- 20:15 walk to restaurant
20:15 -- 22:00 dinner in De Mezzanine