Personnel
Overall Objectives
Research Program
Application Domains
Highlights of the Year
New Software and Platforms
New Results
Bilateral Contracts and Grants with Industry
Partnerships and Cooperations
Dissemination
Bibliography
XML PDF e-pub
PDF e-Pub


Section: Partnerships and Cooperations

National Initiatives

ANR-DFG Project SMArT

Participants : Haniel Barbosa, Pascal Fontaine, Stephan Merz, Thomas Sturm.

The SMArT (Satisfiability Modulo Arithmetic Theories) project was funded by ANR-DFG Programmes blancs 2013, a bilateral (French-German) program of Agence Nationale de la Recherche and Deutsche Forschungsgemeinschaft DFG. It started in April 2014 and finished in September 2017. The project gathered members of VeriDis in Nancy and Saarbrücken, and the Systerel company.

The objective of the SMArT project was to provide advanced techniques for arithmetic reasoning beyond linear arithmetic for formal system verification, and particularly for SMT. The results feed back into the implementations of Redlog and veriT, which also serve as experimentation platforms for theories, techniques and methods designed within this project.

More information on the project can be found on http://smart.gforge.inria.fr/.

ANR Project IMPEX

Participants : Souad Kherroubi, Dominique Méry.

The ANR Project IMPEX, within the INS program, started in December 2013 for 4 years. It was coordinated by Dominique Méry, the other partners are IRIT/ENSEIHT, Systerel, Supelec, and Telecom Sud Paris. The work reported here also included a cooperation with Pierre Castéran from LaBRI Bordeaux.

Modeling languages provide techniques and tool support for the design, synthesis, and analysis of the models resulting from a given modeling activity, as part of a system development process. These languages quite successfully focus on the analysis of the designed system, exploiting the semantic power of the underlying modeling language. The semantics of this modeling languages are well understood by its users (in particular the system designers), i.e. the semantics is implicit in the model. In general, modeling languages are not equipped with resources, concepts or entities handling explicitly domain engineering features and characteristics (domain knowledge) underlying the modeled systems. Indeed, the designer has to explicitly handle the knowledge resulting from an analysis of this application domain [61], i.e. explicit semantics. At present, making explicit the domain knowledge inside system design models does not obey any methodological rules validated by practice. The users of modeling languages introduce these domain knowledge features through types, constraints, profiles, etc. Our claim is that ontologies are good candidates for handling explicit domain knowledge. They define domain theories and provide resources for uniquely identifying domain knowledge concepts. Therefore, allowing models to make references to ontologies is a modular solution for models to explicitly handle domain knowledge. Overcoming the absence of explicit semantics expression in the modeling languages used to specify systems models will increase the robustness of the designed system models. Indeed, the axioms and theorems resulting from the ontologies can be used to strengthen the properties of the designed models. The objective [50] is to offer rigorous mechanisms for handling domain knowledge in design models.

ANR Project Formedicis

Participant : Dominique Méry.

The ANR Project Formedicis, within the INS program, started in January 2017 for 4 years. It is coordinated by Bruno d'Augsbourg, the partners are ONERA, IRIT/ENSEIHT, ENAC, and LORIA.

During the last 30 years, the aerospace domain has successfully devised rigorous methods and tools for the development of safe functionally-correct software. During this process, interactive software has received a relatively lower amount of attention. However, Human-System Interactions (HSI) are important for critical systems and especially in aeronautics: new generations of aircraft cockpits make use of sophisticated electronic devices that may be driven by more and more complex software applications. The criticality of these applications require a high degree of assurance for their intended behavior. The report by the French Bureau d'Enquêtes et d'Analyses about the crash of the Rio-Paris flight AF 447 in 2009 pointed out a design issue in the behavior of the Flight Director interface as one of the original causes of the crash.

We believe that part of these issues are due to the lack of a well-defined domain specific “hub” language to represent interactive software design in a way that allows system designers to iterate on their designs before injecting them in a development process, and system developers to verify their software against the chosen design. Formedicis aims at designing such a formal hub language L, in which designers can express their requirements concerning the interactive behavior that must be embedded inside the interactive applications. The project will also develop a framework for validating, verifying, and implementing critical interactive applications designed and denoted in L.

More information on the project is available at http://www.agence-nationale-recherche.fr/Project-ANR-16-CE25-0007.

ANR Project PARDI

Participants : Marie Duflot-Kremer, Stephan Merz.

PARDI (Verification of parameterized distributed systems) is funded by ANR. The project started in January 2017 for a duration of 48 months. The project partners other than VeriDis are Toulouse INP (coordinator), Université Paris Sud, and Université Paris Marie Curie.

Distributed systems and algorithms are parameterized by the number of participating processes, the communication model, the fault model, and more generally the properties of interaction among the processes. The project aims at providing methodological and tool support for verifying parameterized systems, using combinations of model checking and theorem proving. VeriDis contributes its expertise on TLA+ and its verification tools, and the integration with the Cubicle model checker is a specific goal of the project.

More information on the project is available at http://pardi.enseeiht.fr/.

Inria IPL HAC SPECIS

Participants : Marie Duflot-Kremer, Stephan Merz.

The goal of the HAC SPECIS (High-performance Application and Computers: Studying PErformance and Correctness In Simulation) project is to answer methodological needs of HPC application and runtime developers and to allow studying real HPC systems with respect to both correctness and performance. To this end, this Inria Project Lab assembles experts from the HPC, formal verification, and performance evaluation communities.

HAC SPECIS started in 2016. VeriDis contributes through its expertise in formal verification techniques. In particular, our goal is to extend the functionalities of exhaustive and statistical model checking within the SimGrid platform.

Inria Technological Development Action CUIC

Participants : Jasmin Christian Blanchette, Simon Cruanes.

Most “theorems” initially given to a proof assistant are incorrect, whether because of a typo, a missing assumption, or a fundamental flaw. Novices and experts alike can enter invalid formulas and find themselves wasting hours, or even days, on an impossible proof. This project, funded by Inria and running from 2015 to 2017, supported the development of a counterexample generator for higher-order logic. This new tool, called Nunchaku, is intended for integration with various proof assistants. The project was coordinated by Jasmin Blanchette and also involved Inria Saclay – Île de France (Toccata group) and Inria Rennes – Bretagne Atlantique (Celtique group), among others. Simon Cruanes worked on Nunchaku from October 2015 to September 2017, whereas Blanchette has developed an Isabelle frontend. Four releases have taken place so far, and the tool is an integral part of the Isabelle2017 official release. Work has started on Coq and TLAPS frontends, and we will soon work on a Lean frontend as well. The tool is described in [62] and was presented at a workshop last year [57]. A noteworthy development this year is the creation of a backend called SMBC, based on new ideas by Cruanes about how to combine SAT solving and narrowing [29].