Abstract
We propose a convex-optimization-based framework for computation of invariant measures of polynomial dynamical systems and Markov processes, in discrete and continuous time. The set of all invariant measures is characterized as the feasible set of an infinite-dimensional linear program (LP). The objective functional of this LP is then used to single out a specific measure (or a class of measures) extremal with respect to the selected functional such as physical measures, ergodic measures, atomic measures (corresponding to, e.g., periodic orbits) or measures absolutely continuous w.r.t. to a given measure. The infinite-dimensional LP is then approximated using a standard hierarchy of finite-dimensional semidefinite programming problems, the solutions of which are truncated moment sequences, which are then used to reconstruct the measure. In particular, we show how to approximate the support of the measure as well as how to construct a sequence of weakly converging absolutely continuous approximations. As a by-product, we present a simple method to certify the nonexistence of an invariant measure, which is an important question in the theory of Markov processes. The presented framework, where a convex functional is minimized or maximized among all invariant measures, can be seen as a generalization of and a computational method to carry out the so-called ergodic optimization, where linear functionals are optimized over the set of invariant measures. Finally, we also describe how the presented framework can be adapted to compute eigenmeasures of the Perron–Frobenius operator.
Similar content being viewed by others
Notes
A set of the form (9) is called basic semialgebraic; this class of sets is very rich, including balls, boxes, ellipsoids, discrete sets and various convex and non-convex shapes.
More precisely, F is assumed to be lower semi-continuous with respect to the product topology on the space of sequences \(\mathbb {R}^\infty \). This is in particular satisfied if F depends only on finitely many moments as, for example, in (8).
For a relation of this form of a Markov process to the one specified by the transition kernel, see, e.g., Hernández-Lerma and Lasserre (1996).
References
Bochi, J.: Ergodic optimization of Birkhoff averages and Lyapunov exponents. In: Proceedings of the International Congress of Mathematicians (2018)
Bollt, E.M.: The path towards a longer life: on invariant sets and the escape time landscape. Int. J. Bifurc. Chaos 15(05), 1615–1624 (2005)
Chernyshenko, S.I., Goulart, P., Huang, D., Papachristodoulou, A.: Polynomial sum of squares in fluid dynamics: a review with a look ahead. Philos. Trans. R. Soc. A: Math. Phys. Eng. Sci. 372(2020), 20130350 (2014)
Cross, W.P., Romeijn, H.E., Smith, R.L.: Approximating extreme points of infinite dimensional convex sets. Math. Oper. Res. 23(2), 433–442 (1998)
Fantuzzi, G., Goluskin, D., Huang, D., Chernyshenko, S.I.: Bounds for deterministic and stochastic dynamical systems using sum-of-squares optimization. SIAM J. Appl. Dyn. Syst. 15(4), 1962–1988 (2016)
Fazel, M.: Matrix rank minimization with applications. Ph.D. Thesis, Electrical Engineering Department, Stanford University (2002)
Gaitsgory, V., Quincampoix, M.: Linear programming approach to deterministic infinite horizon optimal control problems with discounting. SIAM J. Control Optim. 48(4), 2480–2512 (2009)
Goluskin, D.: Bounding averages rigorously using semidefinite programming: mean moments of the Lorenz system. J. Nonlinear Sci. 28(2), 621–651 (2017)
Gustafsson, B., Putinar, M., Saff, E.B., Stylianopoulos, N.: Bergman polynomials on an archipelago: estimates, zeros and shape reconstruction. Adv. Math. 222(4), 1405–1460 (2009)
Henrion, D.: Semidefinite characterisation of invariant measures for one-dimensional discrete dynamical systems. Kybernetika 48(6), 1089–1099 (2012)
Henrion, D., Lasserre, J.B., Löfberg, J.: Gloptipoly 3: moments, optimization and semidefinite programming. Optim. Methods Softw. 24, 761–779 (2009)
Hernández-Lerma, O., Lasserre, J.B.: Discrete-Time Markov Control Processes: Basic Optimality Criteria. Springer, Berlin (1996)
Jenkinson, O.: Ergodic optimization. Discrete Contin. Dyn. Syst. 15(1), 197 (2006)
Jenkinson, O.: Ergodic optimization in dynamical systems. Ergod. Theory Dyn. Syst. 39(10), 2593–2618 (2019)
Junge, O., Kevrekidis, I.G.: On the sighting of unicorns: a variational approach to computing invariant sets in dynamical systems. Chaos: Interdiscip. J. Nonlinear Sci. 27(6), 063102 (2017)
Korda, M.: Computing controlled invariant sets from data using convex optimization. SIAM J. Control Optim. 58(5), 2871–2899 (2020)
Korda, M., Henrion, D., Jones, C.N.: Convex computation of the maximum controlled invariant set for polynomial control systems. SIAM J. Control Optim. 52(5), 2944–2969 (2014)
Korda, M., Henrion, D., Jones, C.N.: Controller design and value function approximation for nonlinear dynamical systems. Automatica 67, 54–66 (2016)
Korda, M., Henrion, D., Jones, C.N.: Convergence rates of moment-sum-of-squares hierarchies for optimal control problems. Syst. Control Lett. 100, 1–5 (2017)
Lasota, A., Mackey, M.C.: Chaos, Fractals, and Noise: Stochastic Aspects of Dynamics. Springer, Berlin (1994)
Henrion, Didier, Naldi, Simone, Din, Mohab Safey El: Exact algorithms for linear matrix inequalities. SIAM J. Optim. 26(4), 2512–2539 (2016)
Lasserre, J.B.: Moments, Positive Polynomials and their Applications. Imperial College Press (2010)
Lasserre, J.B.: Global optimization with polynomials and the problem of moments. SIAM J. Optim. 11(3), 796–817 (2001)
Lasserre, J.B., Henrion, D., Prieur, C., Trélat, E.: Nonlinear optimal control via occupation measures and LMI relaxations. SIAM J. Control Optim. 47(4), 1643–1666 (2008)
Lasserre, J.B., Pauwels, E.: The empirical christoffel function in statistics and machine learning. arXiv:1701.02886 (2017)
Löfberg, J.: Yalmip: a toolbox for modeling and optimization in Matlab. In: Proceedings of the IEEE CACSD Conference, Taipei, Taiwan (2004)
Magron, V., Henrion, D., Forets, M.: Semidefinite characterization of invariant measures for polynomial systems. Submitted for publication (2018)
Meyn, S.P., Tweedie, R.L.: Markov Chains and Stochastic Stability. Springer, Berlin (2012)
Mezić, Igor, Banaszuk, Andrzej: Comparison of systems with complex behavior. Physica D 197, 101–133 (2004)
Ozay, N., Lagoa, C., Sznaier, M.: Set membership identification of switched linear systems with known number of subsystems. Automatica 51, 180–191 (2015)
Ozay, N., Sznaier, M., Lagoa, C.: Convex certificates for model (in) validation of switched affine systems with unknown switches. IEEE Trans. Autom. Control 59(11), 2921–2932 (2014)
Pauwels, E., Lasserre, J. B.: Sorting out typicality with the inverse moment matrix sos polynomial. In: Advances in Neural Information Processing Systems (NIPS) (2016)
Phelps, R.R.: Lectures on Choquet’s Theorem. Springer, Berlin (2001)
Putinar, M.: Positive polynomials on compact semi-algebraic sets. Indiana Univ. Math. J. 42, 969–984 (1993)
Schlosser, C., Korda, M.: Converging outer approximations to global attractors using semidefinite programming. arXiv preprint arXiv:2005.03346 (2020)
Sturm, J.: Using SeDuMi 1.02, a Matlab toolbox for optimization over symmetric cones. Optim. Methods Softw. 11:625–653 (1999)
Tobasco, I., Goluskin, D., Doering, C.: Optimal bounds and extremal trajectories for time averages in dynamical systems. APS M1–002 (2017)
Acknowledgements
This work benefited from discussions with Victor Magron. This research was supported in part by the ARO-MURI grant W911NF-17-1-0306. The research of M. Korda was also supported by the Swiss National Science Foundation under grant P2ELP2_165166 and by the Czech Science Foundation (GACR) under contract No. 20-11626Y.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Paul Newton.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Korda, M., Henrion, D. & Mezić, I. Convex Computation of Extremal Invariant Measures of Nonlinear Dynamical Systems and Markov Processes. J Nonlinear Sci 31, 14 (2021). https://doi.org/10.1007/s00332-020-09658-1
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00332-020-09658-1