covid
Buscar en
Revista Iberoamericana de Automática e Informática Industrial RIAI
Toda la web
Inicio Revista Iberoamericana de Automática e Informática Industrial RIAI Control Multimodal en Entornos Inciertos usando Aprendizaje por Refuerzos y Proc...
Información de la revista
Vol. 12. Núm. 4.
Páginas 385-396 (octubre - diciembre 2015)
Compartir
Compartir
Descargar PDF
Más opciones de artículo
Visitas
3056
Vol. 12. Núm. 4.
Páginas 385-396 (octubre - diciembre 2015)
Open Access
Control Multimodal en Entornos Inciertos usando Aprendizaje por Refuerzos y Procesos Gaussianos
Multimodal Control in Uncertain Environments using Reinforcement Learning and Gaussian Processes
Visitas
3056
Mariano De Paulaa,
,1
, Luis O. Ávilab, Carlos Sánchez Reinosoc, Gerardo G. Acostaa
a Núcleo INTELYMEC – CIFICEN – CONICET, Facultad de Ingeniería, Universidad Nacional del Centro de la Provincia de Buenos Aires – UNCPBA, Av. del Valle 5737, Olavarría B7400JWI, Argentina
b Instituto de Desarrollo y Diseño- INGAR (CONICET-UTN), Avellaneda 3657, Santa Fe S3002 GJC, Argentina
c Centro de Diseño y Optimización de Sistemas, Facultad de Tecnología y Ciencias Aplicadas, Universidad Nacional de Catamarca-CONICET, Maximio Victoria 55, San Fernando del Valle de Catamarca 4700 SFV
Este artículo ha recibido

Under a Creative Commons license
Información del artículo
Resumen
Texto completo
Bibliografía
Descargar PDF
Estadísticas
Resumen

El control de sistemas complejos puede ser realizado descomponiendo la tarea de control en una secuencia de modos de control, o simplemente modos. Cada modo implementa una ley de retroalimentación hasta que se activa una condición de terminación, en respuesta a la ocurrencia de un evento exógeno/endógeno que indica que la ejecución del modo debe finalizar. En este trabajo se presenta una propuesta novedosa para encontrar una política de conmutación óptima para resolver el problema de control optimizando alguna medida de costo/beneficio. Una política óptima implementa un programa de control multimodal óptimo, el cual consiste en un encadenamiento de modos de control. La propuesta realizada incluye el desarrollo y formulación de un algoritmo basado en la idea de la programación dinámica integrando procesos Gaussianos y aprendizaje Bayesiano activo. Mediante el enfoque propuesto es posible realizar un uso eficiente de los datos para mejorar la exploración de las soluciones sobre espacios de estados continuos. Un caso de estudio representativo es abordado para demostrar el desempeño del algoritmo propuesto.

Palabras clave:
Control multimodal
Programación dinámica
Procesos Gaussianos
Incertidumbre
Política
Abstract

The control of complex systems can be done decomposing the control task into a sequence of control modes, or modes for short. Each mode implements a parameterized feedback law until a termination condition is activated in response to the occurrence of an exogenous/endogenous event, which indicates that the execution mode must end. This paper presents a novel approach to find an optimal switching policy to solve a control problem by optimizing some measure of cost/benefit. An optimal policy implements an optimal multimodal control program, consisting in a sequence of control modes. The proposal includes the development of an algorithm based on the idea of dynamic programming integrating Gaussian processes and Bayesian active learning. In addition, an efficient use of the data to improve the exploration of the continuous state spaces solutions can be achieved through this approach. A representative case study is discussed and analyzed to demonstrate the performance of the proposed algorithm.

Keywords:
Multimodal Control
Dynamic Programming
Gaussian Processes
Uncertainty
Policy
Referencias
[Abate et al., 2008]
Abate, Alessandro, Maria Prandini, John Lygeros, Shankar Sastry.
Probabilistic reachability and safety for controlled discrete time stochastic hybrid systems.
Automatica, 44 (2008), pp. 2724-2734
[Adamek and Sobotka, 2008]
Adamek, F., M Sobotka, O Stursberg. 2008. Stochastic optimal control for hybrid systems with uncertain discrete dynamics. Proceedings of the IEEE International Conference on Automation Science and Engineering, 23-28. Washington D.C.
[Åström and Karl Johan, 2003]
Åström, Karl Johan, Bo Bernhardsson. 2003. System with Lebesgue Sampling. Directions in Mathematical Systems Theory and Optimization, LNCIS 268. LNCIS. Springer-Verlag Berlin Heidelberg.
[Axelsson et al., 2007]
H. Axelsson, Y. Wardi, M. Egerstedt, E.I. Verriest.
Gradient descent approach to optimal mode scheduling in hybrid dynamical systems.
Journal of Optimization Theory and Applications, 136 (2007), pp. 167-186
[Azuma et al., 2010]
Azuma, Shun-ichi, Jun-ichi Imura, Toshiharu Sugie.
Lebesgue piecewise affine approximation of nonlinear systems.
Nonlinear Analysis: Hybrid Systems, 4 (2010), pp. 92-102
[Barton et al., 2006]
Barton, I. Paul, Cha Kun Lee, Mehmet Yunt.
Optimization of hybrid systems.
Computers & Chemical Engineering, 30 (2006), pp. 1576-1589
[Bemporad et al., 2011]
A. Bemporad, S. Di, Cairano.
Model-predictive control of discrete hybrid stochastic automata.
IEEE Transactions on Automatic Control, 56 (2011), pp. 1307-1321
[Bemporad et al., 1999]
Bemporad, Alberto, Manfred Morari.
Control of systems integrating logic, dynamics, and constraints.
Automatica, 35 (1999), pp. 407-427
[Bensoussan and Menaldi, 2000]
Bensoussan, A.,J. L. Menaldi. 2000. Stochastic hybrid control. Journal of Mathematical Analysis and Applications 249.
[Bertsekas and Dimitri, 2000]
Bertsekas, Dimitri P. 2000. Dynamic Programming and Optimal Control, Vol. I. 2nd ed. Athena Scientific.
[Blackmore et al., 2010]
L. Blackmore, M. Ono, A. Bektassov, B.C. Williams.
A probabilistic particle-control approximation of chance-constrained stochastic predictive control.
IEEE Transactions on Robotics, 26 (2010), pp. 502-517
[Borrelli et al., 2005]
Borrelli, Francesco, Mato Baotić, Alberto Bemporad, Manfred Morari.
Dynamic programming for constrained optimal control of discrete-time linear hybrid systems.
Automatica, 41 (2005), pp. 1709-1721
[Bryson et al., 1975]
Bryson, Jr Arthur E., Yu-Chi Ho. 1975. Applied optimal control: optimization, estimation and control. Revised. Taylor & Francis.
[Busoniu et al., 2010]
Busoniu, Lucian, Robert Babuska, Bart De Schutter,Damien Ernst. 2010. Reinforcement learning and dynamic programming using function approximators. 1.a ed. CRC Press.
[Cassandras et al., 2007]
Cassandras, Christos G., John Lygeros. 2007. Stochastic hybrid systems. Boca Raton: Taylor & Francis.
[Deisenroth and Marc Peter, 2010]
Deisenroth, Marc Peter. 2010. Efficient Reinforcement Learning Using Gaussian Processes. KIT Scientific Publishing.
[Deisenroth et al., 2009]
Deisenroth, Marc Peter, Carl Edward Rasmussen, Jan Peters.
Gaussian process dynamic programming.
Neurocomputing, 72 (2009), pp. 1508-1524
[Di Cairano et al., 2009]
S. Di Cairano, A. Bemporad, J. Júlvez.
Event-driven optimization-based control of hybrid systems with integral continuous-time dynamics.
Automatica, 45 (2009), pp. 1243-1251
[Ding et al., 2009]
X.-C. Ding, Y. Wardi, M. Egerstedt.
On-line optimization of switched-mode dynamical systems.
IEEE Transactions on Automatic Control, 54 (2009), pp. 2266-2271
[Egerstedt et al., 2006]
M. Egerstedt, Y. Wardi, H. Axelsson.
Transition-Time Optimization for Switched-Mode Dynamical Systems.
IEEE Transactions on Automatic Control, 51 (2006), pp. 110-115
[Girard, 2004]
Girard, Agathe. 2004. Approximate methods for propagation of uncertainty with gaussian process models. University of Glasgow.
[Kuss, 2006]
Kuss, M. 2006. Gaussian process models for robust regression, classification, and reinforcement learning. Technische Universite Darmstadt.
[Liberzon, 2003]
Liberzon, Daniel. 2003. Switching in systems and control. Systems & Control: Foundations & Applications. Boston: Birkhäuser Boston Inc.
[Lincoln and Rantzer, 2006]
B. Lincoln, A. Rantzer.
Relaxing Dynamic Programming.
IEEE Transactions on Automatic Control, 51 (2006), pp. 1249-1260
[Lunze et al., 2010]
Lunze, Jan, Daniel Lehmann.
A state-feedback approach to event-based control.
Automatica, 46 (2010), pp. 211-215
[Mehta, 2005]
Mehta, Tejas,Magnus Egerstedt. 2005. Learning multi-modal control programs. Hybrid Systems: Computation and Control, 466-479. Lecture Notes in Computer Science. Springer Berlin.
[Mehta et al., 2006]
Mehta, R. Tejas, Magnus Egerstedt.
An optimal control approach to mode generation in hybrid systems.
Nonlinear Analysis, 65 (2006), pp. 963-983
[Mehta et al., 2008]
Mehta, R. Tejas, Magnus Egerstedt.
Multi-modal control using adaptive motion description languages.
Automatica, 44 (2008), pp. 1912-1917
[Pajares Martin-Sanz and De la Cruz Garcia, 2010]
Pajares Martin-Sanz, G., y De la Cruz Garcia J.M. 2010. Aprendizaje automático. Un enfoque práctico, Cap. 12, Aprendizaje por Refuerzos. RA-MA.
[Rantzer, 2006]
A. Rantzer.
Relaxed Dynamic Programming in Switching Systems.
Control Theory and Applications IEE Proceedings -, 153 (2006), pp. 567-574
[Rasmussen, 2006]
Rasmussen, Carl Edward,Christopher K. I. Williams. 2006. Gaussian processes for machine learning. MIT Press.
[Rosenstein et al., 2004]
Rosenstein, Michael T.,Andrew G. Barto. 2004. Supervised Actor-Critic Reinforcement Learning. Handbook of Learning and Approximate Dynamic Programming, 359-380. John Wiley & Sons, Inc.
[Salichs et al., 2010]
M.A. Salichs, M. Malfaz, J.F. Gorostiza.
Toma de Decisiones en Robótica.
Revista Iberoamericana de Automática e Informática Industrial RIAI, 7 (2010), pp. 5-16
[Shi et al., 2006]
Shi, Peng, G.P. Yuanqing Xia, D. Liu, Rees.
On designing of sliding-mode control for stochastic jump systems.
IEEE Transactions on Automatic Control, 51 (2006), pp. 97-103
[Song et al., 2010]
Song, Chunyue, Ping Li.
Near optimal control for a class of stochastic hybrid systems.
Automatica, 46 (2010), pp. 1553-1557
[Sutton et al., 1998]
Sutton, Richard S.,Andrew G. Barto. 1998. Reinforcement learning: An introduction. MIT Press.
[Verdinelli et al., 1992]
Verdinelli, Isabella, B. Joseph, Kadane.
Bayesian designs for maximizing information and outcome.
Journal of the American Statistical Association, 87 (1992), pp. 510-515
[Xu et al., 2003]
Xu, Xuping,Panos J. Antsaklis. 2003. Results and perspectives on computational methods for optimal control of switched systems. Proceedings of the 6th international conference on Hybrid systems: computation and control, 540-555. Springer-Verlag.
[Xu et al., 2011]
Xu, Yan-Kai, Xi-Ren Cao.
Lebesgue-Sampling-Based Optimal Control Problems with Time Aggregation.
IEEE Transactions on Automatic Control, 56 (2011), pp. 1097-1109
[Zhang et al., 2009]
Zhang, Wei, A. Jianghai Hu, Abate.
On the value functions of the discrete-time switched LQR problem.
IEEE Transactions on Automatic Control, 54 (2009), pp. 2669-2674

URL: www.fio.unicen.edu.ar.

Opciones de artículo