A linear programming methodology for approximate dynamic programming

• Autorzy: Díaz Henry , Sala Antonio , Armesto Leopoldo

Identyfikatory

DOI

10.34768/amcs-2020-0028

• Języki publikacji: EN

Abstrakty

EN

The linear programming (LP) approach to solve the Bellman equation in dynamic programming is a well-known option for finite state and input spaces to obtain an exact solution. However, with function approximation or continuous state spaces, refinements are necessary. This paper presents a methodology to make approximate dynamic programming via LP work in practical control applications with continuous state and input spaces. There are some guidelines on data and regressor choices needed to obtain meaningful and well-conditioned value function estimates. The work discusses the introduction of terminal ingredients and computation of lower and upper bounds of the value function. An experimental inverted-pendulum application will be used to illustrate the proposal and carry out a suitable comparative analysis with alternative options in the literature.

Słowa kluczowe

EN

linear programming; approximate dynamic programming; control applications; neural network

PL

programowanie liniowe; programowanie dynamiczne; sieć neuronowa

• Wydawca: Oficyna Wydawnicza Uniwersytetu Zielonogórskiego
• Czasopismo: International Journal of Applied Mathematics and Computer Science
• Rocznik: 2020
• Tom: Vol. 30, no. 2
• Strony: 363--375
• Opis fizyczny: Bibliogr. 29 poz., rys., wykr.

Twórcy

autor

Díaz Henry

• University Institute of Industrial Control Systems and Computing, Polytechnic University of Valencia, Camino de Vera, s/n, 46022 València, Spain

autor

Sala Antonio

• University Institute of Industrial Control Systems and Computing, Polytechnic University of Valencia, Camino de Vera, s/n, 46022 València, Spain

autor

Armesto Leopoldo

• Design and Manufacturing Institute, Polytechnic University of Valencia, Camino de Vera, s/n, 46022 València, Spain

Bibliografia

• [1] Allgower, F. and Zheng, A. (2012). Nonlinear Model Predictive Control, Springer, New York, NY.
• [2] Ariño, C., Querol, A. and Sala, A. (2017). Shape-independent model predictive control for Takagi–Sugeno fuzzy systems, Engineering Applications of Artificial Intelligence 65(1): 493–505.
• [3] Armesto, L., Girbés,V., Sala, A., Zima, M. and Šmídl, V. (2015). Duality-based nonlinear quadratic control: Application to mobile robot trajectory-following, IEEE Transactions on Control Systems Technology 23(4): 1494–1504.
• [4] Armesto, L., Moura, J., Ivan, V., Erden, M.S., Sala, A. and Vijayakumar, S. (2018). Constraint-aware learning of policies by demonstration, International Journal of Robotics Research 37(13–14): 1673–1689.
• [5] Bertsekas, D.P. (2017). Dynamic Programming and Optimal Control, Vol. 1, 4th Edn, Athena Scientific, Belmont, MA.
• [6] Bertsekas, D.P. (2019). Reinforcement Learning and Optimal Control, Athena Scientific, Belmont, MA.
• [7] Busoniu, L., Babuska, R., De Schutter, B. and Ernst, D. (2010). Reinforcement Learning and Dynamic Programming Using Function Approximators, CRC Press, Boca Raton, FL.
• [8] Cervellera, C., Wen, A. and Chen, V.C. (2007). Neural network and regression spline value function approximations for stochastic dynamic programming, Computers & Operations Research 34(1): 70–90.
• [9] De Farias, D.P. and Van Roy, B. (2003). The linear programming approach to approximate dynamic programming, Operations Research 51(6): 850–865.
• [10] Deisenroth, M.P., Neumann, G. and Peters, J. (2013). A survey on policy search for robotics, Foundations and Trends in Robotics 2(1–2): 1–142.
• [11] Díaz, H., Armesto, L. and Sala, A. (2019). Metodología de programación dinámica aproximada para control óptimo basada en datos, Revista Iberoamericana de Automática e Informática industrial 16(3): 273–283.
• [12] Díaz, H., Armesto, L. and Sala, A. (2020). Fitted Q-function control methodology based on Takagi–Sugeno systems, IEEE Transactions on Control Systems Technology 28(2): 477–488.
• [13] Lagoudakis, M.G. and Parr, R. (2003). Least-squares policy iteration, Journal of Machine Learning Research 4(Dec): 1107–1149.
• [14] Lewis, F.L. and Liu, D. (2013). Reinforcement Learning and Approximate Dynamic Programming for Feedback Control, Wiley, Hoboken, NJ.
• [15] Lewis, F.L. and Vrabie, D. (2009). Reinforcement learning and adaptive dynamic programming for feedback control, IEEE Circuits and Systems Magazine 9(3): 32–50.
• [16] Lewis, F., Vrabie, D. and Syrmos, V. (2012). Optimal Control, 3rd Edn, John Wiley & Sons, Hoboken, NJ.
• [17] Liu, D., Wei, Q., Wang, D., Yang, X. and Li, H. (2017). Adaptive Dynamic Programming with Applications in Optimal Control, Springer, Berlin.
• [18] Manne, A.S. (1960). Linear programming and sequential decisions, Management Science 6(3): 259–267.
• [19] Marsh, L.C. and Cormier, D.R. (2001). Spline Regression Models, Number 137, Sage, Thousand Oaks, CA.
• [20] Munos, R., Baird, L.C. and Moore, A.W. (1999). Gradient descent approaches to neural-net-based solutions of the Hamilton–Jacobi–Bellman equation, International Joint Conference on Neural Networks, Washington, DC, USA, Vol. 3, pp. 2152–2157.
• [21] Munos, R. and Szepesvári, C. (2008). Finite-time bounds for fitted value iteration, Journal of Machine Learning Research 9(May): 815–857.
• [22] Powell, W.B. (2011). Approximate Dynamic Programming: Solving the Curses of Dimensionality, 2nd Edn, Wiley, Hoboken, NJ.
• [23] Preitl, S., Precup, R.-E., Preitl, Z., Vaivoda, S., Kilyeni, S. and Tar, J.K. (2007). Iterative feedback and learning control. servo systems applications, IFAC Proceedings Volumes 40(8): 16–27.
• [24] Rantzer, A. (2006). Relaxed dynamic programming in switching systems, IEE Proceedings: Control Theory and Applications 153(5): 567–574.
• [25] Robles, R., Sala, A. and Bernal, M. (2019). Performance-oriented quasi-LPV modeling of nonlinear systems, International Journal of Robust and Nonlinear Control 29(5): 1230–1248.
• [26] Sutton, R.S. and Barto, A.G. (2018). Reinforcement Learning: An Introduction, 2nd Edn, MIT Press, Cambridge, MA.
• [27] Tan, K., Zhao, S. and Xu, J. (2007). Online automatic tuning of a proportional integral derivative controller based on an iterative learning control approach, IET Control Theory Applications 1(1): 90–96.
• [28] Zajdel, R. (2013). Epoch-incremental reinforcement learning algorithms, International Journal of Applied Mathematics and Computer Science 23(3): 623–635, DOI: 10.2478/amcs-2013-0047.
• [29] Zhao, D., Liu, J., Wu, R., Cheng, D. and Tang, X. (2019). An active exploration method for data efficient reinforcement learning, International Journal of Applied Mathematics and Computer Science 29(2): 351–362, DOI: 10.2478/amcs-2019-0026.

Uwagi

PL

Opracowanie rekordu ze środków MNiSW, umowa Nr 461252 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2020).