Réunion MODEL - Séminaire Renato Vizuete

Titre : Consensus on open multi-agent systems over graphs sampled from graphons
Résumé : Open multi-agent systems are characterized by arrivals, departures, or replacements of agents during the evolution of the dynamics. These types of systems are better approximations of modern networks, where agents constantly enter and leave the systems, modifying the behavior of the group (e.g., social networks, epidemics, vehicle platoons, swarms of robots). Due to the complexity of the system, most works focus mainly on the agent internal dynamics and the processes for arrivals and departures, neglecting the influence of the network dynamics (changes on the set of nodes and connections). In this case, a promising mathematical object for incorporating the effect of changes in the network is the concept of graph functions (graphons) which are limits of sequences of dense graphs. In this talk, we present a framework based on graphons for the analysis of open multi-agent systems where arrivals, departures and replacements are determined by stochastic processes. In the first part, we propose the disagreement as a descriptor to analyze the system and we derive an upper bound for its asymptotic behavior in expectation. In the second part, we exploit properties of graphons to obtain an upper bound that depends only on the graphon and is independent of the number of agents.
Biogaphie : Renato Vizuete is an Assistant Professor at Université de Lorraine, France. He received the M.S. degree (très bien) in Systems, Control, and Information Technologies from Université Grenoble Alpes, France, in 2019, and the PhD degree in Automatic Control from Université Paris-Saclay, France, in 2022. He was a FNRS postdoctoral researcher (CR) at UCLouvain, Belgium, until 2026. His research interests include optimization, control theory, machine learning, network theory and system identification.

le lien suivant : https://teams.microsoft.com/meet/347316762766726?p=ISwP1fN0SCOTuYqPNz.