Modular multi-agent reinforcement learning of linked multi-component robotic systems

  1. FERNANDEZ GAUNA, BORJA
Supervised by:
  1. Manuel Graña Romay Director

Defence university: Universidad del País Vasco - Euskal Herriko Unibertsitatea

Fecha de defensa: 23 April 2012

Committee:
  1. Ángel Pascual del Pobil Ferré Chair
  2. Francisco Xabier Albizuri Irigoyen Secretary
  3. Bruno Apolloni-Ghetti Committee member
  4. Michal Wozniak Committee member
  5. Richard J. Duro Fernández Committee member

Type: Thesis

Teseo: 115276 DIALNET

Abstract

THE CONTENTS OF THIS THESIS CAN BE SUMMARIZED AS TWO MAIN IDEAS: MODULAR TECHNIQUES TO DECOMPOSE A REINFORCEMENT LEARNING TASK IN OVER-CONSTRAINED ENVIRONMENTS SUCH AS LINKED-MCRS AS SEVERAL CONCURRENT SUB-TASKS, AND EXTENSION OF THESE MODULAR REINFORCEMENT LEARNING APPROACHES TO MULTI-AGENT REINFORCEMENT LEARNING ENVIRONMENTS.