Solving Continuous Action/State Problem in Q-Learning Using Extended Rule Based Fuzzy Inference System

  • Kim, Min-Soeng (Department of Electrical Engineering and Computer Science, KAIST) ;
  • Lee, Ju-Jang (Department of Computer Science and Electrical Engineering, KAIST)
  • Published : 2001.09.01

Abstract

Q-learning is a kind of reinforcement learning where the agent solves the given task based on rewards received from the environment. Most research done in the field of Q-learning has focused on discrete domains, although the environment with which the agent must interact is generally continuous. Thus we need to devise some methods that enable Q-learning to be applicable to the continuous problem domain. In this paper, an extended fuzzy rule is proposed so that it can incorporate Q-learning. The interpolation technique, which is widely used in memory-based learning, is adopted to represent the appropriate Q value for current state and action pair in each extended fuzzy rule. The resulting structure based on the fuzzy inference system has the capability of solving the continuous state about the environment. The effectiveness of the proposed structure is shown through simulation on the cart-pole system.

Keywords

References

  1. Machine Learning v.8 Technical note: Q-learning C. J. C. H. Watkins;P. Dayan
  2. Proc. FUZZ-IEEE'96, 5th Int. Conf. Fuzzy Systems. Fuzzy interpolation based Q-learning with continuous states and action T. Horiuchi;A. Fujino;O. Katai;T. Sawaragi
  3. Proc. IROS 97 Region-based Q-learning using convex clustering approach J. H. Kim;I. H. Suh;S. R. Oh;Y. J. Cho;Y. K. Chung
  4. CONIS Technical Report 96-088, University of Massachusetts Experiments with reinforcement learning in problems with continuous state and action spaces J. C. Santamaria;R. S. Sutton;A. Ram
  5. Proc. Herrera Delayed reinforcement, fuzzy {Q}-learning and fuzzy logic controllers A. Bonarini
  6. Proc. FUZZ-IEEE'94, 3rd Int. Conf. Fuzzy Systems. v.1 Fuzzy Q-learning and dynamical fuzzy Q-learning P. Y. Glorennec
  7. SICE Annual Conference of Japan(International session) Fuzzy inference-based reinforcement learning of dynamic recurrent neural networks K.B. Sim
  8. Reinforcement Learning: An Introduction Richard S. Sutton;Andrew G. Barto
  9. Machine Learning v.22 Reinforcement learning with replacing eligibility traces S. P. Singh;R. S. Sutton
  10. IEEE Transactions on Fuzzy Systems v.2 no.1 Reinforcement structure / parameter learning for neural-network-based fuzzy logic control systems C. T. Lin;C. S. G. Lee
  11. Proceedings of the ICMT' 99 Self-organizing modular neural network for reinforcement learning problem S. G. Hong;M. S. Kim;W. Kim;J. J. Lee