A Michigan style architecture for learning finite state controllers: a first step - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2004

A Michigan style architecture for learning finite state controllers: a first step

Samuel Landau
  • Fonction : Auteur
Olivier Sigaud

Résumé

In this paper, we describe FACS, a new Michigan style architecture able to build Finite-State Automata controllers for agents learning to solve non-Markov problems. FACS relies on a population of partial automata and implements a Reinforcement Learning algorithm to compute the strength of each automaton and a Genetic Algorithm to select and discover efficient automata. We detail our approach and present very preliminary results.
Fichier non déposé

Dates et versions

hal-01498518 , version 1 (30-03-2017)

Identifiants

  • HAL Id : hal-01498518 , version 1

Citer

Samuel Landau, Olivier Sigaud. A Michigan style architecture for learning finite state controllers: a first step. 7th International Workshop on Learning Classifier Systems, Jun 2004, Seattle, WA, United States. ⟨hal-01498518⟩
54 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More