Gaël Glorian

Senior R&D Engineer · Huawei Technologies France, Paris Research Centre · Boulogne-Billancourt, Île-de-France, France
gael.glorian@huawei.com

Gaël Glorian is a Senior R&D Engineer in the Constraint Programming team of the Paris Research Centre at Huawei Technologies France since September 2021. Before joining Huawei he did is PhD at CRIL, Artois University in France on Hybridization of Clause Learning Techniques in Constraint Programming. He has been interested in improving the CSP (Constraint Satisfaction Problem) solving techniques, especially in a hybrid context using the power of SAT (Boolean Satisfaction Problem) inference engines. From this work, he has developed a tool called Nacre that won two gold medals at the XCSP3 Competition - Minitrack in 2018 and 2019. After his PhD, he worked as a postdoctoral researcher at Ubisoft Bordeaux and LaBRI, Bordeaux University for the project KIWI of Région Nouvelle-Aquitaine to meet technological challenges on a set of fields related to automation of scalable interactive simulations in real time and to analyse actors and their behavior. His missions were to improve procedural content generation techniques and to formally guarantee certain properties at the end of this type of generation.


Education

Ph.D in Artificial Intelligence

CRIL · Centre de Recherche en Informatique de Lens · Faculté Jean Perrin · Lens, Hauts-de-France · France
Title: Hybridization of clause learning techniques in constraint programming [ bib | http | pdf ]
Supervisors: Frédéric Boussemart, Jean-Marie Lagniez, Christophe Lecoutre and Bertrand Mazure
Keywords: constraint programming, SAT, clauses, nogoods, hybridization

Abstract: This thesis belongs to the field of constraint programming (CP), one of the most efficient paradigms for solving many problems (of a combinatorial nature) in AI. We have been interested in improving the CSP (constraint satisfaction problem) solving techniques, especially in a hybrid context using the power of SAT (Boolean satisfaction problem) inference engines. We have developed several conflict analysis techniques to extract useful information for resolution, through fine analysis followed by a form of learning. To achieve this goal, we first strengthened the power of nogoods recorded by the resolution system by proposing several rules to combine them. In addition, we integrated a SAT resolution engine into a CP system based on the notion of lazy explanations. The use of a SAT engine is motivated by their high efficiency in producing and manipulating simple forms of nogoods in clausal form. The NACRE software, a generic reasoning engine, is the result of this work; In particular, it was designed to be a hybrid solver, solving constraint satisfaction problems using dedicated or SAT-inspired methods. Our generic (i.e., valid for any problem) approach has proved very effective in practice (NACRE has won 2 gold medals at the XCSP competitions 2018 and 2019). In order to enrich the SAT / CP hybridization, we conducted a study on the quality of clauses produced by the SAT engine. This allowed us to propose new methods for clauses database reduction, minimization and new search heuristics.

See on theses.fr See on CRIL website
2016 · 2019

Master’s degree in computer science

Faculté Jean Perrin · Lens, Hauts-de-France · France
2014 · 2016

Bachelor’s degree in computer science

Faculté Jean Perrin · Lens, Hauts-de-France · France
2009 · 2014

Scientific Baccalaureate

Lycée Louis Pasteur · Hénin-Beaumont, Hauts-de-France · France
2006 · 2009

Publications

International

The Dungeon Variations Problem Using Constraint Programming.bib | http | pdf]
Gaël Glorian, Adrien Debesson, Sylvain Yvon-Paliot, and Laurent Simon.
In Laurent D. Michel, editor, 27th International Conference on Principles and Practice of Constraint Programming, CP 2021, Montpellier, France (Virtual Conference), October 25-29, 2021, volume 210 of LIPIcs, pages 27:1--27:16. Schloss Dagstuhl - Leibniz-Zentrum für Informatik, 2021.

NACRE - A nogood and clause reasoning engine.bib | http | pdf]
Gael Glorian, Jean-Marie Lagniez, and Christophe Lecoutre.
In Elvira Albert and Laura Kovács, editors, LPAR 2020: 23rd International Conference on Logic for Programming, Artificial Intelligence and Reasoning, Alicante, Spain, May 22-27, 2020, volume 73 of EPiC Series in Computing, pages 249--259. EasyChair, 2020.

Hybridization of clause learning techniques in constraint programming.bib | http | pdf ]
Gael Glorian.
Theses, Université d'Artois, December 2019.

pFactory: A generic library for designing parallel solvers.bib | http | pdf]
Gilles Audemard, Gael Glorian, Jean-Marie Lagniez, Valentin Montmirail, and Nicolas Szczepanski.
In Hans Weghorn, editor, the 16th International Conference on Applied Computing, AC'19, pages 89--96, 2019.

An incremental sat-based approach to the graph colouring problem.bib | http | pdf]
Gael Glorian, Jean-Marie Lagniez, Valentin Montmirail, and Nicolas Szczepanski.
In Thomas Schiex and Simon de Givry, editors, Principles and Practice of Constraint Programming · 25th International Conference, CP 2019, Stamford, CT, USA, September 30 · October 4, 2019, Proceedings, volume 11802 of Lecture Notes in Computer Science, pages 213--231. Springer, 2019.

An incremental sat-based approach to reason efficiently on qualitative constraint networks.bib | http | pdf]
Gael Glorian, Jean-Marie Lagniez, Valentin Montmirail, and Michael Sioutis.
In John N. Hooker, editor, Principles and Practice of Constraint Programming · 24th International Conference, CP 2018, Lille, France, August 27-31, 2018, Proceedings, volume 11008 of Lecture Notes in Computer Science, pages 160--178. Springer, 2018.

Combining nogoods in restart-based search.bib | http | pdf]
Gael Glorian, Frédéric Boussemart, Jean-Marie Lagniez, Christophe Lecoutre, and Bertrand Mazure.
In J. Christopher Beck, editor, Principles and Practice of Constraint Programming · 23rd International Conference, CP 2017, Melbourne, VIC, Australia, August 28 · September 1, 2017, Proceedings, volume 10416 of Lecture Notes in Computer Science, pages 129--138. Springer, 2017.


National

Génération de donjons à l'aide de la programmation par contraintes.bib | pdf]
Gael Glorian, Adrien Debesson, Sylvain Yvon-Paliot, and Laurent Simon.
In Seizièmes Journées Francophones de Programmation par Contraintes, JFPC 2021, Nice, France, 22 au 24 juin, 2021, Actes.

Une approche sat incrémentale pour raisonner efficacement sur les réseaux de contraintes qualitatives.bib | pdf]
Gael Glorian, Jean-Marie Lagniez, Valentin Montmirail, and Michael Sioutis.
In Quizièmes journées Francophones de Programmation par Contraintes, JFPC 2019, Albi, France, 10 au 13 juin, 2019, Actes, volume 15, pages 67--76. AFPC, 2019.

Combinaison de nogoods extraits au redémarrage.bib | pdf]
Gael Glorian, Frédéric Boussemart, Jean-Marie Lagniez, Christophe Lecoutre, and Bertrand Mazure.
In Treizièmes journées Francophones de Programmation par Contraintes, JFPC 2017, Montreuil-sur-mer, France, 13 au 15 juin, 2017, Actes, volume 13, pages 55--64. AFPC, 2017.


Solver description

Nacre.bib | http | pdf]
Gael Glorian.
In Christophe Lecoutre and Olivier Roussel, editors, Proceedings of the 2018 XCSP3 Competition, pages 85--85, 2019.


Tools

NACRE

NACRE, for Nogood And Clause Reasoning Engine, is a constraint solver written in C++. It is based on a modular architecture designed to work with generic constraints while implementing several state-of-the-art search methods and heuristics. Interestingly, its data structures have been carefully designed to play around nogoods and clauses, making it suit-able for implementing learning strategies. NACRE was submitted to theCSP MiniTrack of the 2018 and 2019 XCSP3 [8] competitions where it took the first place. This paper gives a general description of NACRE as a framework. We present its kernel, the available search algorithms, and the default settings (notably, used for XCSP3 competitions), which makes NACRE efficient in practice when used as a black-box solver.

pFactory

pFactory is a parallel library designed to support and facilitate the implementation of parallel solvers in C++. It provides robust implementations of parallel algorithms and allows seamlessly sharing mechanisms, divide-and-conquer or portfolio methods. pFactory is not related to a specific problem and can very easily be incorporated in order to solve any kind of combinatorial problem (SAT, CSP, MAXSAT...).

Awards

Activities, Comitees & Supervising

Supervising

Suruthy Sekar · Master’s student (M2) @ Paris Dauphine · End-of-studies Internship · 2022 · Huawei · SAT/CP Synergies
Hugo Martin · Master’s student (M2) @ Bordeaux INP · End-of-studies Internship · 2021 · LaBRI & Ubisoft · Procedural Generation Of Narrative Content
Quentin Sauvage · Master’s student (M1) @ Artois university · 2019 · CRIL · Incremental Debugging Tool for CP Solvers
Ouahab Fenniche · Master’s student (M1) @ Artois university · 2019 · CRIL · Outil pour Exécuter et Analyser des Expérimentations

Program Committee

Seizièmes journées Francophones de Programmation par Contraintes · JFPC'21 · Nice, France
Seizièmes journées Francophones de Programmation par Contraintes · JFPC'20 · Postponed to 2021 (COVID-19) · Nice, France
Quizièmes journées Francophones de Programmation par Contraintes · JFPC'19 · Albi, France

Invited Talk

Dungeon Variations Using Constraint Programming · Ubisoft · La Forge Open House · Worldwide
An introduction to Constraint Programming · LaBRI · Bordeaux, France
Constraint Programming and Real-World Applications · Ubisoft Bordeaux · Bordeaux, France
Solveurs et Hybridation SAT/CP - De MAC à un moteur générique de raisonnement de clauses en CP · I3S · Sophia Antipolis, France

Administration

Hcéres · Pre-selected as "young PhD" expert · 2020-2023
Qualificationas a lecturer · Section 27, Computer Science · 2020-2024
Elected member of the laboratory council · CRIL · 2018-2020

Organizing Committee

Treizièmes journées Francophones de Programmation par Contraintes · JFPC'17· Montreuil sur Mer, France
Journées des Doctorants · CRIL · 2017 · Le Touquet, France
The 30th International Conference on Industrial, Engineering, Other Applications of Applied Intelligent Systems · IEA/AIE 2017 · Arras, France
The 24th International Conference on Principles and Practice of Constraint Programming · CP'18 · Lille, France

Teaching

Networks · M1 · Faculté Jean Perrin · Lens, France
Advanced C Programming · L3 · Faculté Jean Perrin · Lens, France
Programming & Algorithm 5 · L3 · Faculté Jean Perrin · Lens, France
Programming & Algorithm 4 · L2 · Faculté Jean Perrin · Lens, France
Digital Culture · L1 · Faculté Jean Perrin · Lens, France