DMCA
Heterogeneous and hierarchical cooperative learning via combining decision trees (2006)
| Venue: | in Proc. of the IEEE/RSJ Conf. on Intelligent Robots and Systems [IROS], (Beijing |
| Citations: | 2 - 2 self |
BibTeX
@INPROCEEDINGS{Asadpour06heterogeneousand,
author = {Masoud Asadpour and Majid Nili Ahmadabadi},
title = {Heterogeneous and hierarchical cooperative learning via combining decision trees},
booktitle = {in Proc. of the IEEE/RSJ Conf. on Intelligent Robots and Systems [IROS], (Beijing},
year = {2006}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Abstract — Decision trees, being human readable and hierarchically structured, provide a suitable mean to derive state-space abstraction and simplify the inclusion of the available knowledge for a Reinforcement Learning (RL) agent. In this paper, we address two approaches to combine and purify the available knowledge in the abstraction trees, stored among different RL agents in a multi-agent system, or among the decision trees learned by the same agent using different methods. Simulation results in nondeterministic football learning task provide strong evidences for enhancement in convergence rate and policy performance. I.
Keyphrases
decision tree hierarchical cooperative learning available knowledge state-space abstraction multi-agent system abstract decision tree nondeterministic football learning task convergence rate reinforcement learning suitable mean abstraction tree strong evidence human readable different rl agent different method policy performance simulation result
