An active learning method for solving competitive multi-agent decision-making and control problems

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Fabiani, Filippo, Bemporad, Alberto
Natura: Preprint
Pubblicazione: 2022
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913536136970240
author Fabiani, Filippo
Bemporad, Alberto
author_facet Fabiani, Filippo
Bemporad, Alberto
contents To identify a stationary action profile for a population of competitive agents, each executing private strategies, we introduce a novel active-learning scheme where a centralized external observer (or entity) can probe the agents' reactions and recursively update simple local parametric estimates of the action-reaction mappings. Under very general working assumptions (not even assuming that a stationary profile exists), sufficient conditions are established to assess the asymptotic properties of the proposed active learning methodology so that, if the parameters characterizing the action-reaction mappings converge, a stationary action profile is achieved. Such conditions hence act also as certificates for the existence of such a profile. Extensive numerical simulations involving typical competitive multi-agent control and decision-making problems illustrate the practical effectiveness of the proposed learning-based approach.
format Preprint
id arxiv_https___arxiv_org_abs_2212_12561
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle An active learning method for solving competitive multi-agent decision-making and control problems
Fabiani, Filippo
Bemporad, Alberto
Systems and Control
Machine Learning
Multiagent Systems
Optimization and Control
To identify a stationary action profile for a population of competitive agents, each executing private strategies, we introduce a novel active-learning scheme where a centralized external observer (or entity) can probe the agents' reactions and recursively update simple local parametric estimates of the action-reaction mappings. Under very general working assumptions (not even assuming that a stationary profile exists), sufficient conditions are established to assess the asymptotic properties of the proposed active learning methodology so that, if the parameters characterizing the action-reaction mappings converge, a stationary action profile is achieved. Such conditions hence act also as certificates for the existence of such a profile. Extensive numerical simulations involving typical competitive multi-agent control and decision-making problems illustrate the practical effectiveness of the proposed learning-based approach.
title An active learning method for solving competitive multi-agent decision-making and control problems
topic Systems and Control
Machine Learning
Multiagent Systems
Optimization and Control
url https://arxiv.org/abs/2212.12561