Baptiste Buchi

🇬🇧 Hi ! 🇬🇧

I’m Baptiste Buchi, I have a PhD in computer science specializing in the use of artificial intelligence in multi-agent systems, distributed systems, and the reconfiguration of modular robots.

My thesis, “Learning System for Self-Reconfiguration of Micro-Robot Networks,” was carried out in the OMNI team at the FEMTO-ST laboratory and was completed in 2023. I have since worked at LORIA in the LARSEN team and am currently a lecturer and researcher at CESI in the LINEACT laboratory.

🇫🇷 Bonjour ! 🇫🇷

Je suis Baptiste Buchi, docteur en informatique spécialisé en utilisation d’intelligence artificielle sur des systèmes multi-agents, les systèmes distribués et la reconfiguration de robots modulaires.

Ma thèse “Système d’apprentissage pour le problème d’auto-reconfiguration des réseaux de micro-robots” a été effectué dans l’équipe OMNI du laboratoire FEMTO-ST et s’est conclu en 2023. J’ai depuis travaillé au LORIA dans l’équipe LARSEN et suis actuellement Enseignant-chercheur à CESI dans le laboratoire LINEACT.

selected publications

Machine learning for modular robots self-reconfiguration problem

Baptiste Buchi, Hakim Mabed, Frédéric Lassabe, and 1 more author

Journal of Ambient Intelligence and Humanized Computing, Mar 2025

DOI
Translation based Self Reconfiguration Algorithm for 6-lattice Modular Robots

Baptiste Buchi, Hakim Mabed, Frédéric Lassabe, and 2 more authors

In 2021 20th International Symposium on Parallel and Distributed Computing (ISPDC), Mar 2021

DOI
Mixed integer quadratic programming for modular robots self reconfiguration problem

Hakim Mabed, Hervé Manier, Marie-Ange Manier, and 1 more author

Operational Research, Mar 2026

Abs DOI

Recent years have shown continuous progress in nanoscale assembly with the development of affordable large-scale manufacturing technologies. Modular robotics and, especially, the programmable matter is one of the emanations of this evolution, which consists of a set of identical modular robots capable of reconfiguring into any shape. However, the success of such a technology depends on the provided algorithmic solutions that allow shifting from the current configuration to the targeted shape. Major works concerning the modular robots self-reconfiguration problem (MRSR) are based on local and/or distributed heuristics that do not guarantee the optimality of the reconfiguration proposals in terms of energy and convergence time. Furthermore, centralized approaches employ constraint relaxation to reduce the combinatorial complexity of the problem. Therefore, obtained solutions are not practically usable in the physical world. In this paper, we provide the first modeling of the modular-robots self-reconfiguration problem into Mixed-Integer Quadratic Programming Model (MIQP). The model includes at the best of our knowledge all the physical, synchronization, and operational constraints of the horizontal 6-lattice MRSR problem. Modeling the MRSR problem as a MIQP problem allows to use exact optimization solvers to pre-calculate the best reconfiguration plan or to prove the problem’s unfeasibility. The obtained results show the relevance of the MIQP approach, especially for small and mid-size instances. Problems with 10 robots could be solved in a few seconds with sometimes the proof of optimality.