Evaluation of the improved penalty avoiding rational policy making algorithm in real world environment

Authors:
Kazuteru Miyazaki;Masaki Itou;Hiroaki Kobayashi
Affiliations:
National Institution for Academic Degrees and University Evaluation, Japan;Toshiba Tec Corporation, Japan;Meiji University, Japan
Venue:
ACIIDS'12 Proceedings of the 4th Asian conference on Intelligent Information and Database Systems - Volume Part I
Year:
2012

Citing 5
Cited 0

Introduction to Reinforcement Learning

Introduction to Reinforcement Learning
An Analysis of Actor/Critic Algorithms Using Eligibility Traces: Reinforcement Learning with Imperfect Value Function

ICML '98 Proceedings of the Fifteenth International Conference on Machine Learning
Algorithms for Inverse Reinforcement Learning

ICML '00 Proceedings of the Seventeenth International Conference on Machine Learning
Exploration and apprenticeship learning in reinforcement learning

ICML '05 Proceedings of the 22nd international conference on Machine learning
Introduction of fixed mode states into online profit sharing and its application to waist trajectory generation of biped robot

EWRL'11 Proceedings of the 9th European conference on Recent Advances in Reinforcement Learning

Quantified Score

Hi-index	0.00

Visualization

Abstract

We focus on a potential capability of Exploitation-oriented Learning (XoL) in non-Markov multi-agent environments. XoL has some degree of rationality in non-Markov environments and is also confirmed the effectiveness by computer simulations. Penalty Avoiding Rational Policy Making algorithm (PARP) that is one of XoL methods was planed to learn a penalty avoiding policy. PARP is improved to save memories and to cope with uncertainties, that is called Improved PARP. Though the effectiveness of Improved PARP has been confirmed on computer simulations, there is no result in real world environment. In this paper, we show the effectiveness of Improved PARP in real world environment using a keepaway task that is a testbed of multi-agent soccer environment.