Abstract
We propose Extended Q-learning. To accommodate continuous state space directly and to improve its generalization capability. Through EQ-learning, an action-value function is represented by the summation of weighted base functions, and an autonomous robot adjusts weights of base functions at learning stage. Other parameters (center coordinates, variance and so on) are adjusted at unification stage where two similar functions are unified to a simpler function.
Chapter PDF
Similar content being viewed by others
References
M. Asada, S. Noda and K. Hosoda: Action Based Sensor Space Segmentation For Soccer Robot Learning. Applied Artificial Intelligence, Vol.12, No.2-3 (1998) 149–164.
Digney, B. L.: Learning Hierarchical Control Structures for Multiple Tasks and Changing Environments. From animals to animats 5: SAB 98,(1998).
Andrew G. Barto, Richard S. Sutton, and Charles W. Anderson: Neuronlike Adaptive Elements that can solve difficult learning control problems. (1983) 834–846.
Hiroaki Kitano, Minoru Asada, Yasuo Kuniyoshi, Itsuki Noda, Eiichi Osawa, and Hitoshi Matsubara;: RoboCup: A Challenge Problem for AI and Robotics. Lecture Notes in Artificial Intelligence Vol.1395 (1998)1–19.
Norihiko Ono, Yoshihiro Fukuta: Learning Coordinated Behavior in a Continuous Environment. Lecture Notes in Artificial Intelligence Vol.1221 (1997) 73–81.
C.J.C.H. Watkins: Learning from delayed rewards. PhD thesis, University of Cambridge (1989).
Whitehead:A complexity analysis of cooperative mechanisms in reinforcement learning. Proc. AAAI-91 (1991) 607–613.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2001 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Enokida, S., Ohasi, T., Yoshida, T., Ejima, T. (2001). Extended Q-Learning: Reinforcement Learning Using Self-Organized State Space. In: Stone, P., Balch, T., Kraetzschmar, G. (eds) RoboCup 2000: Robot Soccer World Cup IV. RoboCup 2000. Lecture Notes in Computer Science(), vol 2019. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-45324-5_11
Download citation
DOI: https://doi.org/10.1007/3-540-45324-5_11
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-42185-6
Online ISBN: 978-3-540-45324-6
eBook Packages: Springer Book Archive