Abstract
Reinforcement Learning is a learning methodology through an agent interacting with the environment. Actor-Critic methods have a separated memory structure to present the independence between the policy and the value function. In systems where the states are defined as continuous, there is a problem with dimensionality, and an approximation method has to be used. The classic Reinforcement Learning algorithms can be combined with Fuzzy Logic techniques to store all value functions, since Fuzzy Logic has been proved to be an effective universal approximator. This work propose a Fuzzy Actor-Critic method to compute and store the state values using fuzzy logic to get a state approximation. Phenol is one of the most important water pollutants on chemistry industries. A phenol biodegradation process consist on a Sequence Batch Reactor (SBR), that need an near optimal filling policy for its correct operation. Fuzzy Actor Critic learning strategy offers a operation policy and it can be linguistically interpreted by the process experts, this approach can be useful to propose a comprehensive filling policy of a biodegradation SBR process.
Keywords
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
Bertsekas, D., Tsitsiklis, J.: Neuro-Dynamic Programming. Athena Scientific, Belmont (1996)
Sutton, R.: Learning to predict by the method of temporal differences. Machine Learning 3, 9–44 (1988)
Castro, J.: Fuzzy logic controllers are universal aproximators. IEEE Trans. on Systems, Man and Cybernetics 25, 629–635 (1995)
García, J.A., Ramírez, C., Ben-Youssef, C., Waissman, J., Vázquez, G.A.: Modelado de la biodegradación de fenol por lodos activados aclimatados. Revista Internacional de Contaminación Ambiental 21, 802–807 (2004)
Vázquez, G., Ben Youssef, C., Waissman, J.: Two-step modeling of the biodegradation of phenol by an acclimated activated sludge. Chemical Engineering Journal 117, 245–252 (2006)
Glorennec, P.Y., Jouffe, G.: Fuzzy Q–learning. In: Proceedings of Fuzz-IEEE, Sixth International Conference on Fuzzy Systems, Barcelona, pp. 659–662 (1997)
Duan, Y., Xu, X.: Fuzzy reinforcement learning and its application in robot navigation. In: Proceedings of the Fourth International Conference on Machine Learning and Cybernetics, pp. 899–904 (2005)
Vengerov, E., Berenj, H.: A convergent actor–critic-based FRL algorithm with application to power management of wireless transmitters. IEEE Transactions on Fuzzy Systems 11, 478–485 (2003)
Vengerov, E., Bambos, N., Berenj, H.: A fuzzy reinforcement learning approach to power control in wireless transmitters. IEEE Trans. on Systems, Man and Cybernetics 35, 765–778 (2005)
Wang, X., Cheng, Y., Yi, J.: A fuzzy actor–critic reinforcement learning network. Information Sciences 178, 3764–3781 (2007)
Barto, A., Sutton, R.S., Anderson, C.W.: Neuronlike elements that can solve difficult learning control problems. IEEE Trans. on Systems, Man and Cybernetics 13, 1038–1044 (1983)
Sutton, R., Barto, A.: Reinforcement Learning, an Introduction. MIT Press, Cambridge (1998)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2008 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Franco Flores, E., Waissman Vilanova, J., García Lamont, J. (2008). Learning the Filling Policy of a Biodegradation Process by Fuzzy Actor–Critic Learning Methodology. In: Gelbukh, A., Morales, E.F. (eds) MICAI 2008: Advances in Artificial Intelligence. MICAI 2008. Lecture Notes in Computer Science(), vol 5317. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-88636-5_23
Download citation
DOI: https://doi.org/10.1007/978-3-540-88636-5_23
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-88635-8
Online ISBN: 978-3-540-88636-5
eBook Packages: Computer ScienceComputer Science (R0)