Abstract
In this paper, we propose a method to reduce relearning costs by ap-plying successful policies to estimated failing states in reinforcement learning. When an environment is changed to another, relearning is needed in order to acquire an appropriate policy in the new environment. In order to reduce re-learning costs, an algorithm has been proposed to estimate failing states using a decision tree C4.5, and it relearns new policies only for the estimated failing states in the new environment. We try to reduce failing states furthermore by applying successful policies to the estimated failing states. Computer simulations show that our method can reduce relearning costs and improve the successful rate in reinforcement learning.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Sutton, R.S., Barto, A.G.: Reinforcement Learning: An Introduction. MIT Press, Cambridge (1998)
Kaelbling, L.P., Littman, M.L., Moore, A.W.: Reinforcement learning: A survey. Journal of Artificial Intelligence Research 4, 237–285 (1996)
Minato, T., Asada, M.: Environmental change adaptation for mobile robot navigation. Journal of the Robotics Society of Japan 18(5), 706–712 (2000)
Shen, W.M.: Discovery as autonomous learning from the environment, Machine Learning 28, 143–156 (1993)
Matsui, T., Inuzuka, N., Seki, H.: Adapting to subsequent changes of environment by learning policy preconditions. International Journal of Computer and Information Science 3(1) (2002)
Quinlan, J.R.: C4.5: Programs for Machine Learning. 1(1), 81–106 (1986)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Murata, T., Matsumoto, H. (2004). Use of Successful Policies to Relearn for Induced States of Failure in Reinforcement Learning. In: Negoita, M.G., Howlett, R.J., Jain, L.C. (eds) Knowledge-Based Intelligent Information and Engineering Systems. KES 2004. Lecture Notes in Computer Science(), vol 3213. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-30132-5_151
Download citation
DOI: https://doi.org/10.1007/978-3-540-30132-5_151
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-23318-3
Online ISBN: 978-3-540-30132-5
eBook Packages: Springer Book Archive