The Arts et Métiers campus Arts et Métiers invites you to the thesis defense of Quentin Brilhault, doctoral student at LISPEN, on December 6, 2023.
Thesis title
Digital continuity based on learning transformation models: a model-driven approach to interoperability.
Summary
The digital transformation of industry is forcing companies to improve digital continuity within their organizations by strengthening the ability of their distributed systems to communicate and coordinate their activities in order to perform complex tasks such as intelligent planning, anomaly detection, and reducing the ecological footprint of industrial processes.
In an unstable and highly heterogeneous digital environment, constantly evolving both technically with the emergence of new technologies and organizationally with increased demands for agility and adaptability in production chains, establishing and maintaining interoperability between systems is a crucial task.
Model transformations, the cornerstone of Model-Driven Architecture, could provide a concrete solution to the requirements of dynamic and sustainable interoperability between systems.
This is especially true given that a new paradigm, using machine learning techniques, could simplify the creation and maintenance of transformation models by automatically learning the rules of transformation between models.
Two contributions are presented in this thesis:
- An application of reinforcement learning principles to automatically derive transformation rules;
- An experimental protocol to evaluate and validate the ability of the proposed approach to automatically infer transformation models while complying with Industry 4.0 specifications.
A benchmark of existing approaches concludes this work, demonstrating the effectiveness of reinforcement learning techniques in learning transformation rules that link two different metamodels.
Composition of the jury
- Mr. Julien Le Duigou, University Professor, Roberval UTC (rapporteur)
- Mr. Hervé Panetto, University Professor, CRAN, Telecom Nancy (rapporteur)
- Mr. Vincent Cheutet, University Professor, DIS-lab, INSA Lyon (examiner)
- Ms. Esma Yahia, co-supervisor, Associate Professor, LISPEN, Arts et Métiers examiner)
- Mr. Lionel Roucoules, thesis supervisor, University Professor, LISPEN, Arts et Métiers examiner)
- Dr. Françoise Perrel, CNES (examiner)
Practical information
Location: Arts et Métiers Campus Arts et Métiers (lecture hall J001)
Date: December 6, 2023
Time: 10:00 a.m.
Link to follow the defense via Teams.
