Learning Lyapunov terminal costs from data for complexity reduction in nonlinear model predictive control. (English) Zbl 07900070
Summary: A classic way to design a nonlinear model predictive control (NMPC) scheme with guaranteed stability is to incorporate a terminal cost and a terminal constraint into the problem formulation. While a long prediction horizon is often desirable to obtain a large domain of attraction and good closed-loop performance, the related computational burden can hinder its real-time deployment. In this article, we propose an NMPC scheme with prediction horizon \(N=1\) and no terminal constraint to drastically decrease the numerical complexity without significantly impacting closed-loop stability and performance. This is attained by constructing a suitable terminal cost from data that estimates the cost-to-go of a given NMPC scheme with long prediction horizon. We demonstrate the advantages of the proposed control scheme in two benchmark control problems.
© 2024 John Wiley & Sons Ltd.
© 2024 John Wiley & Sons Ltd.
MSC:
| 93B45 | Model predictive control |
| 93C10 | Nonlinear systems in control theory |
| 93C55 | Discrete-time control/observation systems |
| 93B11 | System structure simplification |
Keywords:
constrained systems; data-driven control; neural networks; nonlinear model predictive controlReferences:
| [1] | GrüneL, PannekJ. Nonlinear Model Predictive Control. 2nd ed.Springer International Publishing; 2017. |
| [2] | MichalskaH, MayneD. Receding horizon control of nonlinear systems. Proceedings of the 28th IEEE Conference on Decision and Control. Vol 1. IEEE; 1989:107‐108. |
| [3] | RawlingsJ, MayneD, DiehlM. Model Predictive Control: Theory, Computation and Design. Nob Hill Publishing; 2019. |
| [4] | ChenH, AllgöwerF. A quasi‐infinite horizon nonlinear model predictive control scheme with guaranteed stability. Automatica. 1998;34(10):1205‐1217. · Zbl 0947.93013 |
| [5] | ChenS, SaulnierK, AtanasovN, et al. Approximating explicit model predictive control using constrained neural networks. 2018 Annual American Control Conference (ACC). IEEE; 2018:1520‐1527. |
| [6] | HertneckM, KöhlerJ, TrimpeS, AllgöwerF. Learning an approximate model predictive controller with guarantees. IEEE Control Syst Lett. 2018;2(3):543‐548. |
| [7] | LuciaS, KargB. A deep learning‐based approach to robust nonlinear model predictive control. IFAC‐PapersOnLine. 2018;51:511‐516. |
| [8] | GersnoviezA, BroxM, BaturoneI. High‐speed and low‐cost implementation of explicit model predictive controllers. IEEE Trans Control Syst Technol. 2019;27(2):647‐662. |
| [9] | Nematollah ChangiziKS, SiahiM. Complexity reduction of explicit MPC based on fuzzy reshaped polyhedrons for use in industrial controllers. Int J Syst Sci. 2023;54(3):463‐477. · Zbl 1520.93132 |
| [10] | GrancharovaA, JohansenT. Nonlinear model predictive control. Explicit Nonlinear Model Predictive Control: Theory and Applications. Springer; 2012:43‐66. · Zbl 1401.93004 |
| [11] | CagienardR, GriederP, KerriganE, MorariM. Move blocking strategies in receding horizon control. J Process Control. 2007;17(6):563‐570. |
| [12] | ShekharR, ManzieC. Optimal move blocking strategies for model predictive control. Automatica. 2015;61:27‐34. · Zbl 1327.93166 |
| [13] | MakarowA, RösmannC, BertramT. Suboptimal nonlinear model predictive control with input move‐blocking. Int J Control. 2022;97:450‐459. · Zbl 1537.93210 |
| [14] | MuehlebachM, D’AndreaR. A method for reducing the complexity of model predictive control in robotics applications. IEEE Robot Autom Lett. 2019;4(3):2516‐2523. |
| [15] | LawrynczukM. Nonlinear model predictive control for processes with complex dynamics: a parameterisation approach using Laguerre functions. Int J Appl Math Comput Sci. 2020;30(1):35‐46. · Zbl 1461.93128 |
| [16] | PanG, FaulwasserT. NMPC in active subspaces: dimensionality reduction with recursive feasibility guarantees. Automatica. 2023;147:110708. · Zbl 1505.93057 |
| [17] | BoggioM, NovaraC, TaragnaM. Nonlinear model predictive control: an optimal search domain reduction. IFAC‐PapersOnLine. 2023;56(2):6253‐6258. |
| [18] | OhtsukaT. A continuation/GMRES method for fast computation of nonlinear receding horizon control. Automatica. 2004;40:563‐574. · Zbl 1168.93340 |
| [19] | DiehlM, FerreauH, HaverbekeN. Efficient numerical methods for nonlinear MPC and moving horizon estimation. In: MagniL (ed.), RaimondoDM (ed.), AllgöwerF (ed.), eds. Nonlinear Model Predictive Control. Lecture Notes in Control and Information Sciences. Springer; 2009:391‐417. · Zbl 1195.93038 |
| [20] | LimonD, AlamoT, SalasF, CamachoE. On the stability of constrained MPC without terminal constraint. IEEE Trans Automat Contr. 2006;51(5):832‐836. · Zbl 1366.93483 |
| [21] | GieslP, HafsteinS. Review on computational methods for Lyapunov functions. Discrete Continuous Dyn Syst Ser B. 2015;20:2291‐2331. · Zbl 1337.37001 |
| [22] | MittalM, GallieriM, QuaglinoA, SalehianS, KoutnikJ. Neural Lyapunov model predictive control. arXiv preprint arXiv:abs/2002.10451, 2020. |
| [23] | Moreno‐MoraF, BeckenbachL, StreifS. Predictive control with learning‐based terminal costs using approximate value iteration. arXiv preprint arXiv:abs/2212.00361, 2022. |
| [24] | AbdufattokhovS, ZanonM, BemporadA. Learning convex terminal costs for complexity reduction in MPC. 2021 60th IEEE Conference on Decision and Control (CDC). IEEE; 2021:2163‐2168. |
| [25] | TuranE, MdoeZ, JäschkeJ. Learning convex objectives to reduce the complexity of model predictive control. arXiv preprint arXiv:abs/2312.02650, 2023. |
| [26] | ZiegerT, SavchenkoA, OehlschlägelT, FindeisenR. Towards safe neural network supported model predictive control. IFAC‐PapersOnLine. 2020;53:5246‐5251. |
| [27] | KargB, LuciaS. Efficient representation and approximation of model predictive control laws via deep learning. IEEE Trans Cybern. 2020;50(9):3866‐3878. |
| [28] | RajhansC, PatwardhanS, PillaiH. Discrete time formulation of quasi infinite horizon nonlinear model predictive control scheme with guaranteed stability. IFAC‐PapersOnLine. 2017;50:7181‐7186. |
| [29] | MayneD, RawlingsJ, RaoC, ScokaertP. Constrained model predictive control: stability and optimality. Automatica. 2000;36(6):789‐814. · Zbl 0949.93003 |
| [30] | ProkhorovD, FeldkampL. Application of support vector machines to Lyapunov function approximation. International Joint Conference on Neural Networks Proceedings (Cat No. 99CH36339). Vol 1. IEEE; 1999:383‐387. |
| [31] | SerpenG. Empirical approximation for Lyapunov functions with artificial neural nets. Proceedings of the 2005 IEEE International Joint Conference on Neural Networks. Vol 2. IEEE; 2005:735‐740. |
| [32] | GabyN, ZhangF, YeX. Lyapunov‐net: a deep neural network architecture for Lyapunov function approximation. arXiv preprint arXiv:abs/2109.13359, 2021. |
| [33] | HornikK, StinchcombeM, WhiteH. Multilayer feedforward networks are universal approximators. Neural Netw. 1989;2:359‐366. · Zbl 1383.92015 |
| [34] | AggarwalC. Neural Networks and Deep Learning: A Textbook. Springer; 2018. · Zbl 1402.68001 |
| [35] | ZangwillW. Nonlinear programming via penalty functions. Manag Sci. 1967;13(5):344‐358. · Zbl 0171.18202 |
| [36] | AnderssonJ, GillisJ, HornG, RawlingsJ, DiehlM. CasADi—a software framework for nonlinear optimization and optimal control. Math Program Comput. 2019;11(1):1‐36. · Zbl 1411.90004 |
| [37] | WächterA, BieglerL. On the implementation of an interior‐point filter line‐search algorithm for large‐scale nonlinear programming. Math Program. 2006;106:25‐57. · Zbl 1134.90542 |
| [38] | VerschuerenR, FrisonG, KouzoupisD, et al. acados: a modular open‐source framework for fast embedded optimal control. Math Program Comput. 2019;14:147‐183. · Zbl 07507324 |
| [39] | PaszkeA, GrossS, MassaF, et al. PyTorch: an imperative style, high‐performance deep learning library. Proceedings of the 33rd International Conference on Neural Information Processing Systems. Curran Associates Inc.; 2019:8024‐8035. |
| [40] | KingmaD, BaJ. Adam: a method for stochastic optimization. arXiv preprint arXiv:abs/1412.6980, 2014. |
| [41] | GoodfellowI, BengioY, CourvilleA. Deep Learning. MIT Press; 2016http://www.deeplearningbook.org · Zbl 1373.68009 |
| [42] | MagniL, deNicolaoG, MagnaniL, ScattoliniR. A stabilizing model‐based predictive control algorithm for nonlinear systems. Automatica. 2001;37:1351‐1362. · Zbl 0995.93033 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
