Abstract
We present a new approach for constructing polytope Lyapunov functions for continuous-time linear switching systems (LSS). This allows us to decide the stability of LSS and to compute the Lyapunov exponent with a good precision in relatively high dimensions. The same technique is also extended for stabilizability of positive systems by evaluating a polytope concave Lyapunov function (“antinorm”) in the cone. The method is based on a suitable discretization of the underlying continuous system and provides both a lower and an upper bound for the Lyapunov exponent. The absolute error in the Lyapunov exponent computation is estimated from above and proved to be linear in the dwell time. The practical efficiency of the new method is demonstrated in several examples and in the list of numerical experiments with randomly generated matrices of dimensions up to 10 (for general linear systems) and up to 100 (for positive systems). The development of the method is based on several theoretical results proved in the paper: the existence of monotone invariant norms and antinorms for positively irreducible systems, the equivalence of all contractive norms for stable systems and the linear convergence theorem.
Similar content being viewed by others
References
F. Amato, R. Ambrosino, M. Ariola, Robust stability via polyhedral Lyapunov functions, in Proceedings of 2009 American Control Conference (ACC2009), 3736–3741 (2009).
H. Alonso, P. Rocha, A general stability test for switched positive systems based on a multidimensional system analysis, IEEE Trans. Automat. Control, 55(11), 2660–2664 (2010).
N.E. Barabanov, An absolute characteristic exponent of a class of linear nonstationary systems of differential equations, Siberian Math. J., 29(4), 521–530 (1988).
N.E. Barabanov, Lyapunov indicator for discrete inclusions, I–III, Automat. Remote Control, 49(2), 152–157 (1988).
N.E. Barabanov, Method for the computation of the Lyapunov exponent of a differential inclusion, Automat. Remote Control, 50(4), 475–479 (1989).
F. Blanchini, S. Miani, Piecewise-linear functions in robust control, Robust control via variable structure and Lyapunov techniques, Lecture Notes in Control and Inform. Sci., 217, Springer, London, 213–243 (1996).
F. Blanchini, S. Miani, A new class of universal Lyapunov functions for the control of uncertain linear systems, IEEE Trans. Automat. Control, 44(3), 641–647 (1999).
F. Blanchini, C. Savorgnan, Stabilizability of switched linear systems does not imply the existence of convex Lyapunov functions, Automatica, 44(4), 1166–1170 (2008).
O.N. Bobyleva, Piecewise-linear Lyapunov functions for linear stationary systems, Automat. Remote Control, 63(4), 540–549 (2002).
L. Chen, New analysis of the sphere covering problems and optimal polytope approximation of convex bodies, J. Approx. Theory, 133(1), 134–145 (2005).
A. Cicone and V.Yu. Protasov. Fast computation of tight bounds for the joint spectral radius, 2013.
M.P. Contzen and J. Raisch, A polytopic approach to switched linear systems. in Proceedings of Control Applications (CCA), IEEE, 1521–1526 (2014).
I. Daubechies, J.C. Lagarias, Sets of matrices all infinite products of which converge. Linear Algebra Appl., 161, 227–263 (1992).
E. De Santis, M.D. Di Benedetto, G. Pola, Stabilizability of linear switching systems, Nonlinear Analysis: Hybrid Systems, 2(3), 750–764 (2008).
R. Edwards, J.J. McDonald and M.J. Tsatsomeros On matrices with common invariant cones with applications in neural and gene networks, Linear Alg. Appl., 398, 37–67 (2005).
L. Fainshil and M. Margaliot, A maximum principle for the stability analysis of positive bilinear control systems with applications to positive linear switched systems, SIAM J. Control Optim., 50(4), 2193–2215 (2012).
L. Fainshil, M. Margaliot, and P. Chigansky, On the stability of positive linear switched systems under arbitrary switching laws, IEEE Trans. Automat. Control, 54(4), 897–899 (2009).
A.F. Filippov, On certain questions in the theory of optimal control, J. SIAM Control Ser. A, 1, 76–84 (1962).
E. Fornasini, M.E. Valcher, Asymptotic stability and stabilizability of special classes of discrete-time positive switched systems, Linear Alg. Appl., 438(4), 1814–1831 (2013).
E. Fornasini, M.E. Valcher, Stability and stabilizability criteria for discrete-time positive switched systems, IEEE Trans. Automat. Control, 57(5), 1208–1221 (2012).
G. Gripenberg, Computing the joint spectral radius, Linear Algebra Appl., 234, 43–60 (1996).
N. Guglielmi, V.Yu. Protasov, Exact computation of joint spectral characteristics of matrices, J. Found. Comput. Math., 13(1), 37–97 (2013).
N. Guglielmi, F. Wirth, and M. Zennaro, Complex polytope extremality results for families of matrices, SIAM J. Matrix Anal. Appl., 27(3), 721–743 (2005).
N. Guglielmi and M. Zennaro, An algorithm for finding extremal polytope norms of matrix families, Linear Alg. Appl., 428(10), 2265–2282 (2008).
N. Guglielmi and M. Zennaro, Finding extremal complex polytope norms for families of real matrices, SIAM J. Matrix Anal. Appl., 31(2), 602–620 (2009).
N. Guglielmi and M. Zennaro, Stability of linear problems: joint spectral radius of sets of matrices, in L. Dieci and N. Guglielmi Eds., Current Challenges in Stability Issues for Numerical Differential Equations. Lecture Notes in Mathematics, Springer, vol. 2082, 265–313 (2014).
L. Gurvits, Stability of discrete linear inclusions, Linear Algebra Appl., 231, 47–85 (1995).
L. Gurvits, R. Shorten, O. Mason, On the stability of switched positive linear systems, IEEE Trans. Automat. Control, 52(6), 1099–1103 (2007).
A.D. Ioffe and V.M. Tihomirov, Theory of extremal problems, Translated from the Russian by Karol Makowski. Studies in Mathematics and its Applications, 6. North-Holland Publishing Co., Amsterdam, New York, 1979.
M. Johansson and A. Rantzer, Computation of piecewise quadratic Lyapunov functions for hybrid systems, IEEE Trans. Automat. Control, 43(4), 555–559 (1998).
P. Julian, J. Guivant, and A. Desages, A parametrization of piecewise linear Lyapunov functions via linear programming, Internat. J. Control, 72(7-8), 702–715 (1999).
R. M. Jungers, The Joint Spectral Radius: Theory and Applications, Vol. 385 in Lecture Notes in Control and Information Sciences, Springer-Verlag. Berlin Heidelberg, (2009).
V.S. Kozyakin, Irreducibility measures in explicit estimation of the joint spectral radius, Diff. Equations and Dyn. Syst., 18(1-2), 91–103 (2010),.
V.S. Kozyakin and A. V. Pokrovsky, The role of controllability-type properties in the study of the stability of desynchronized dynamical systems, Soviet Phys. Dokl., 37(5), 213–215 (1992).
M.G. Krein and M.A. Rutman, Linear operators leaving invariant a cone in a Banach space, Amer. Math. Soc. Transl., 26 (1950), 128 p. [translated from Uspekhi Mat. Nauk 1(23), 3–95 (1948)].
D. Liberzon, Switching in systems and control, Birkhauser, Boston, MA, 2003.
H. Lin and P.J. Antsaklis, Stability and stabilizability of switched linear systems: a survey of recent results, IEEE Trans. Autom. Contr., 54(2), 308–322 (2009).
M. Ludwig, C. Schütt, and E. Werner, Approximation of the Euclidean ball by polytopes, Studia Math., 173(1), 1–18 (2006).
M. Margaliot and G. Langholz, Necessary and sufficient conditions for absolute stability: the case of second-order systems, IEEE Trans. Circuits Systems I Fund. Theory Appl., 50(2), 227–234 (2003).
M. Margaliot and C. Yfoulis, Absolute stability of third-order systems: a numerical algorithm, Automatica J. IFAC, 42(10), 1705–1711 (2006).
S. Marinosson, Lyapunov function construction for ordinary differential equations with linear programming, Dyn. Syst., 17(2), 137–150 (2002).
A.P. Molchanov and E.S. Pyatnitskii, Lyapunov functions, defining necessary and sufficient conditions for the absolute stability of nonlinear nonstationary control systems, Avtomat. Telemekh., 3, 63–73; 4, 5-15; 5, 38–49 (1986).
A.P. Molchanov and E.S. Pyatnitskii, Criteria of asymptotic stability of differential and difference inclusions encountered in control theory, Systems Control Lett., 13(1), 59–64 (1989).
V.I. Opoitsev, Equilibrium and stability in models of collective behaviour, Nauka, Moscow 245 p., (1977)
A. Polanski, Lyapunov function construction by linear programming, IEEE Trans. Autom. Control, 42(7), 1013–1016 (1997).
A. Polanski, On absolute stability analysis by polyhedral Lyapunov functions, Automatica J. IFAC, 36(4), 573–578 (2000).
V. Yu. Protasov, The joint spectral radius and invariant sets of linear operators, Fundamentalnaya i prikladnaya matematika, 2(1), 205–231 (1996).
V.Yu. Protasov, Asymptotic behaviour of the partition function, Sb. Math., 191(3-4), 381–414 (2000).
V.Yu. Protasov, Extremal \(L_p\)-norms of linear operators and self-similar functions, Linear Alg. Appl., 428(10), 2339–2356 (2008).
V.Yu. Protasov, Invariant functionals for random matrices, Functional Anal. Appl., 44(3), 230–233 (2010).
V.Yu. Protasov, When do several linear operators share an invariant cone? Linear Alg. Appl., 433(4), 781–789 (2010).
V.Yu. Protasov and R. Jungers, Is switching systems stability harder for continuous time systems? in Proceedings of 2013 IEEE 52nd Annual Conference on Decision and Control (CDC2013), 704–709 (2013).
G.C. Rota and G. Strang, A note on the joint spectral radius, Kon. Nederl. Acad. Wet. Proc., 63, 379–381 (1960).
L. Rodman, H. Seyalioglu, and I. Spitkovsky, On common invariant cones for families of matrices, Linear Alg. Appl., 432(4), 911–926 (2010).
J. Shen, J. Hu, Stability of discrete-time switched homogeneous systems on cones and conewise homogeneous inclusions, SIAM J. Control Optim., 50(4), 2216–2253 (2012).
J. S. Vandergraft, Spectral properties of matrices which have invariant cones, SIAM J. Appl. Math., 16, 1208–1222 (1968).
H. Wieland, Unzerlegbare, nicht negative Matrizen, Math. Z. 52 (1950), 642–648, and Malbematische WerlcelMalbematical Works, 2, 100–106 de Gruyter, Berlin, 1996.
G. Xie and L. Wang, Controllability and stabilizability of switched linear-systems, Systems and Control Letters, 48(2), 135–155 (2003).
C.A. Yfoulis, Stabilisation of nonlinear systems: the piecewise linear approach, Research Studies, Press Limited, Hertfordshire, England., 2001.
C.A. Yfoulis, Numerical techniques for the absolute stability problem of high-order systems: a conjecture, in Proceedings of the European Control Conference, Kos, Greece, 688–695 (2007).
C.A. Yfoulis, R.A. Shorten, A numerical technique for the stability analysis of linear switched systems, Internat. J. Control, 77(11), 1019–1039 (2004).
Acknowledgments
We are grateful to the referees for their insightful remarks and suggestions. Part of this work was developed during the stay of the third author at the University of L’Aquila under the financial support of INdAM GNCS (Istituto Nazionale di Alta Matematica, Gruppo Nazionale di Calcolo Scientifico) and GSSI (Gran Sasso Science Institute).
The MATLAB routines implementing the construction of the polytope norms and antinorms are available on request to the authors.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Hans Munthe-Kaas.
The research of the first author is supported by Italian INDAM G.N.C.S. The research of the third author is supported by the RFBR Grants Nos. 13-01-00642 and 14-01-00332 and by the Grant of the Dynasty foundation.
Appendix
Appendix
Proof of Theorem 3
It is split into four steps. First we construct a special positively homogeneous monotone convex functional \(\varphi \) and prove (step 1) that it is actually a norm. Then, in step 2, we establish the extremality property of \(\varphi \), i.e., that it is non-increasing on each trajectory. Thus, we have an extremal norm. In steps 3 and 4, we show the existence of a generalized trajectory on the unit sphere starting at an arbitrary point. This is done by considering a special convex optimal control problem and applying the Banach–Alaoglu compactness theorem.
In view of (2), it suffices to consider the case \( \sigma ({\mathcal {A}}) = 0\). We take an arbitrary norm monotone with respect to K, for example, \(\Vert x\Vert = (b,x)\), where \(b \in \mathrm{int}\, K^*\), where \(K^*\) is a dual cone (15). For \(t \ge 0\) and \(z \in K\), denote \(l(z, t) \, = \, \sup \, \{\Vert x(t)\Vert \, , \ A \in {\mathcal {U}}\, [0,t]\, , \, x(0) = z\}\). For every fixed t, the function \(l(\cdot , t)\) is a seminorm on K, i.e., it is positively homogeneous and convex, as a supremum of homogeneous convex functions. Moreover, Corollary 4 implies that l is non-decreasing in z, i.e., if \(z_1 \ge _K z_2\), then \(l(z_1, t) \ge l(z_2, t)\) for each t. The function \(\varphi (x) = \sup \limits _{t \in {\mathbb {R}}_+} l(x, t)\) is, therefore, also a monotone seminorm on K as the supremum of monotone seminorms. Moreover, \(\varphi (x) \ge l(x,0) = \Vert x\Vert \); hence, \(\varphi \) is positive.
Step 1. Let us show that \(\varphi (x) < \infty \) for all x, i.e., \(\varphi \) is a norm on K. Denote by \({\mathcal {L}}\) the set of points \(x \in K\) such that \(\varphi (x) < \infty \). Since \(\varphi \) is convex, homogeneous and monotone on K, it follows that either \({\mathcal {L}}= K\) or \({\mathcal {L}}\) is a face of K. Writing \(\tilde{{\mathcal {L}}}\) for the linear span of \({\mathcal {L}}\), we are going to show that \(A \tilde{{\mathcal {L}}} \subset \tilde{{\mathcal {L}}}\) for each \(A \in {\mathcal {A}}\). If this is not the case, then \(A_0z \notin \tilde{{\mathcal {L}}}\) for some \(z \in {\mathcal {L}}\) and \(A_0 \in {\mathcal {A}}\), hence \((I + tA_0)z \notin \tilde{{\mathcal {L}}}\) for all \(t \in (0, \tau ]\), and therefore \(e^{\, t\, A_0} z \notin \tilde{{\mathcal {L}}}\) for all \(t \in (0, \tau ]\), where \(\tau > 0\) is small enough. Hence, for the control function \(A(t) \equiv A_0, \, \ t \in [0, \tau ]\) and for \(x(0) = z\) we have \(x(\tau ) \notin \tilde{{\mathcal {L}}}\), and consequently, \(\varphi (z) \ge \varphi (x(\tau )) = +\infty \), which is a contradiction. Thus, unless \({\mathcal {L}}= K\), the set \(\tilde{{\mathcal {L}}}\) is a face plane invariant for all operators from \({\mathcal {A}}\). From the K-irreducibility, it follows that either \({\mathcal {L}}= K\) (in which case the proof is completed) or \({\mathcal {L}}= \{0\}\). It remains to show that the latter is impossible.
If \({\mathcal {L}}= \{0\}\), then \(\varphi (z) = +\infty \) for all \(z \in K \setminus \{0\}\). Let \(K_1 = \{z \in K, \, \Vert z\Vert = 1\}\). For every natural n, denote by \({\mathcal {H}}_{\, n}\) the set of points \(z \in K_1\), for which there exists a number \(\tau = \tau (z) \le n\) and a trajectory starting at z such that \(\Vert x(\tau )\Vert > 2\). Since \({\mathcal {H}}_{\, n}\) is open and \(\cup _{n =1}^{\infty }{\mathcal {H}}_{\, n} \, = \, K_1\), from the compactness we conclude that \(\cup _{n = 1}^{N}{\mathcal {U}}_{\, n} \, = \, K_1\) for some natural N. Thus, \(\tau (z) \le N\) for all \(z \in K_1\). Whence, starting from an arbitrary point \(x_0 \in K_1\) one can consequently build a trajectory x(t) and sequences \(\{x_n\}, \{t_n\}\) an \(\{\tau _n\}\) such that \(t_0 = 0, t_n = \sum _{k=0}^{n-1}\tau (x_k)\), \(x_k = x(t_k)\). For this trajectory, \(\Vert x(t_n)\Vert \, > \, 2^n\) and \(t_n \le nN\), hence \(\Vert x(t_n)\Vert \, > \, e^{t_n\ln 2 / N}\). Therefore, \( \sigma ({\mathcal {A}}) \ge \frac{\ln 2}{N} >0\), which contradicts the assumption. Thus, the case \({\mathcal {L}}= \{0\}\) is impossible; hence, \(\varphi \) is a norm.
Step 2. By definition, for any trajectory x(t) the function \(\varphi (x(t))\) is non-increasing in t. Indeed, suppose \(t_1 < t_2\); then \(\varphi (x(t_1))\) is the supremum of \(\Vert y(t)\Vert \, , \, t \in [t_1, +\infty )\) over all possible trajectories \(y(\cdot )\) on the half-line \([t_1, +\infty )\) with the initial condition \(y(t_1) = x(t_1)\). This set of trajectories includes \(x(\cdot )\). Hence this supremum is not smaller than \(\varphi (t_2)\), which is the supremum over a narrower set of trajectories \(y(\cdot )\) on the half-line \([t_2, +\infty )\) with the initial condition \(y(t_2) = x(t_2)\).
Step 3. Thus, we have found a norm \(\Vert x\Vert = \varphi (x)\) which is non-increasing in t on every trajectory x(t). In this norm, the function l(z, t) is non-increasing in t for each \(z \in K\). Hence, the limit \(F(z) = \lim \limits _{t \rightarrow +\infty }l(z, t)\) exists for every \(z \in K\). Let us show that F is a norm we are looking for. First of all, this is a monotone seminorm on K as a limit of monotone seminorms. Second, F(x(t)) is non-increasing in t on every trajectory x(t). Finally, by definition of l(z, t), we have
For every \(\tau > 0\), we denote by \({\mathcal {Q}}_{\, \tau }\) the set of control functions \(A(\cdot ) \in {\mathcal {U}}\, [0, \tau ]\), for which the supremum in the left-hand side of (27) is attained. To show that \({\mathcal {Q}}_{\tau }\) is non-empty, we consider, for an arbitrary \(a \in K^*\), the following optimal control problem:
Since this problem is linear in the control function \(A(\cdot )\), the set \(\mathrm{co}\, ({\mathcal {A}})\) is convex and compact, and the objective function \(\bigl (a , x(\tau ) \bigr )\) is linear, it possesses the optimal solution \((\bar{A}, \bar{x}) \in L_1[0, \tau ] \times W^1_1 [0, \tau ]\) (see, e.g., [18, 29]). Now we take a maximizing sequence \(\{x_i(\cdot )\}_{i = 1}^{\infty }\), for which \(F(x_i(\tau )) \rightarrow F(z)\) in (27) as \(i \rightarrow \infty \). By the compactness, without loss of generality it can be assumed that the sequence \(x_i(z)\) converges to some point \(y \in K\) as \(i \rightarrow \infty \). Taking \(a \in \partial \, F(y)\) (the subdifferential of F at the point y), and solving problem (28) for that a, we obtain \(\bigl (a , \bar{x}(\tau ) \bigr ) = F(z)\), and hence \(F(\bar{x}(\tau )) = F(z)\). Thus, \({\mathcal {Q}}_{\, \tau }\, \ne \, \emptyset \) for each \(\tau > 0\). Note that this set is compact in the weak-* topology of the space \(L_1[0, \tau ]\) due to Banach–Alaoglu theorem. Furthermore, the family \(\{{\mathcal {Q}}_{\, \tau }\}_{\tau > 0}\) is embedded: \({\mathcal {Q}}_{\, \tau _2} \subset {\mathcal {Q}}_{\, \tau _1}\) if \(\tau _2 > \tau _1\). Indeed, if \(\bar{x} \in {\mathcal {Q}}_{\, \tau _2}\), then \(F(\bar{x}(\tau _2)) = F(z)\), and hence \(F(\bar{x}(t))\) equals identically to F(z) on the segment \([0, \tau _2]\). Therefore, it equals identically to F(z) on a smaller segment \([0, \tau _1]\), and so \(\bar{x} \in {\mathcal {Q}}_{\, \tau _1}\). Since an embedded system of non-empty compact sets has a non-empty intersection, it follows that there exists a control function \(\bar{A}\), whose trajectory \(\bar{x} \) with \(\bar{x} (0) = z\) possesses the property \(F(\bar{x} (t)) \equiv F(\bar{x} (0))\, , \ t \in [0, +\infty )\).
Step 4. Thus, we have proved that the seminorm F is invariant: It is non-increasing in t on any trajectory x(t), and for every starting point, there is a trajectory, on which F is identically constant. It remains to show that F is a norm, i.e., F(z) is finite and positive for every \(z \in K\setminus \{0\}\). Since F(z) is defined as a limit of a non-increasing function as \(t \rightarrow +\infty \), we have \(F(z) < \infty \). The positivity is proved by contradiction. Let \({\mathcal {M}}= \{z \in K, \ F(z) = 0\}\). Since F is a seminorm, it follows that \({\mathcal {M}}\) is either entire K or a face of K. If this is a face of K, then as in Step 1 we conclude that its linear span is a common invariant face plane for \({\mathcal {A}}\), which by the irreducibility implies \({\mathcal {M}}= \{0\}\), and the proof is completed. If \({\mathcal {M}}= K\), then \(l(z, t) \rightarrow 0\) as \(t \rightarrow \infty \) for every \(z \in K\). Take an arbitrary \(x_0 \in \mathrm{int}\, K\). There exists a constant \(c > 0\) such that for every \(x \in K\) inequality \(\Vert x\Vert \le c\) implies \(x \, \le _K \, \frac{1}{2}\, x_0\). Let n be such that \(l(x_0, n) < c\). Hence, \(x(n) \, \le _K \, \frac{1}{2}\, x(0)\) for every trajectory x(t) with \(x(0) = x_0\). Applying now Corollary 4 and iterating k times, we get \(x(kn) \, \le _K \, 2^{-k}\, x_0\). Since the norm is monotone, it follows that \(\Vert x(kn)\Vert \, \le \, 2^{-k}\Vert x_0\Vert \, , \, k \in {\mathbb {N}}\). On the other hand, since the norm is non-increasing in t on every trajectory, we see that \(\Vert x(t)\Vert \, \le \, 2^{- \bigl [\frac{t}{n} \bigr ]}\Vert x_0\Vert \), where the brackets denote the integer part. Since for every \(y_0 \in K\) there is a constant C such that \(y_0 \le _K Cx_0\), it follows that for every trajectory y(t) one has \(\Vert y(t)\Vert \, \le \, C 2^{- \bigl [\frac{t}{n} \bigr ]}\Vert y_0\Vert \). Therefore, \( \sigma ({\mathcal {A}}) \le - \frac{\ln 2}{n} < 0\). The contradiction concludes the proof. \(\square \)
Proof of Proposition 6
Replacing the family \({\mathcal {A}}\) by \({\mathcal {A}}+ hI\), where \(h>0\) is large enough, it may be assumed that \(A \ge _K I\) for all \(A \in {\mathcal {A}}\). In this case the family \({\mathcal {B}}\, = \, \{B = A - I \ | \ A \in {\mathcal {A}}\}\) consists of K-irreducible operators that leave K invariant. Take arbitrary vectors \(x \in K \, , \, x^*\in K^* \, , \ \Vert x\Vert = \Vert x^*\Vert = 1\). For every \(B \in {\mathcal {B}}\), we denote \(p_B(x^*, x)\, = \, \max \limits _{n=0, \ldots , d-1}(x^*,B^nx)\). If \(p_B(x^*, x) = 0\), then x is contained in a face of K invariant with respect to B. This contradicts irreducibility of B. Thus, the function p is strictly positive. Hence, by the compactness, there is \(a > 0\) such that \(p_B(x^*, x) \ge a\) for all \(B \in {\mathcal {B}}\) and all \(x \in K , x^*\in K^* \, , \, \Vert x\Vert = \Vert x^*\Vert = 1\). Assume for the moment that the family \({\mathcal {A}}\) is finite: \({\mathcal {A}}= \{A_1, \ldots , A_m\}\) and \(A_k = I +B_k\). Every product \(P = \prod _{k=1}^{md}A_k\) of md operators contains at least d equal terms, say, \(A_i\). Then \(P = \prod _{k=1}^{md}(I+B_k) \ge \sum _{n=0}^{d-1}B_i^n\), and therefore, \((x^*, Px) \, \ge \, p_{B_i}(x^*, x)\, \ge \, a\). In case of general compact set \({\mathcal {A}}\), we take its \(\varepsilon \)-net \({\mathcal {A}}_{\varepsilon } = \{A_i\}_{i=1}^{m(\varepsilon )}\), and to every \(A \in {\mathcal {A}}\), we associate the closets element from \({\mathcal {A}}_{\, \varepsilon }\) (if there are several closest elements, we take any of them). There is a function \(c(\varepsilon )\) such that \(c(\varepsilon ) \rightarrow 0, \varepsilon \rightarrow 0\), and for every product \(\varPi \) of length at most d of operators from \({\mathcal {A}}\), we have \(\Vert \varPi - \varPi '\Vert \le c(\varepsilon )\), where \(\varPi '\) is the corresponding product of operators from \({\mathcal {A}}_{\varepsilon }\). Hence, for every product P of length \(md(\varepsilon )\), we have \((x^*, Px) \, \ge \, a - c(\varepsilon )\). Taking \(\varepsilon \) small enough, so that \(c(\varepsilon ) < a/2\), we see that \((x^*, Px) \, \ge \, a/2\), for every product P of length \(d m(\varepsilon )\) of operators from \({\mathcal {A}}\). Since this holds for all \(x^* \in K^*\, , \, \Vert x^*\Vert = 1\), it follows that \(Px \in \mathrm{int}(K)\) for every \(x \in K\, , \, \Vert x\Vert = 1\). Consequently, the cone PK is embedded in K. Let \(K'\) be the convex hull of all cones PK taken over all products \(P \in {\mathcal {A}}^{\, d m(\varepsilon )}\). The cone \(K'\) is embedded in K and \(AK' \subset K'\, , \ A \in {\mathcal {A}}\); hence, A is \(K'\)-Metzler, which completes the proof. \(\square \)
Proof of Theorem 5
In view of (2), it suffices to consider the case \(\check{\sigma }({\mathcal {A}}) = 0\). Take a positive antinorm on K, for example, \(g(x) = (b,x)\), where \(b \in \mathrm{int}\, K^*\). For \(t \ge 0\) and \(z \in K\), denote \(r(z,t)=\inf \,\{g(x(t)),\,A \in {\mathcal {U}}\, [0,t],\,x(0)=z\}\).
Step 1. For every fixed t, the function \(r(\cdot , t)\) is an antinorm on K. Corollary 4 implies that r is non-decreasing in z. The function \(\psi (x) = \inf \limits _{t \in {\mathbb {R}}_+} r(x, t)\) is, therefore, also an antinorm on K as the infimum of antinorms. This antinorm is non-decreasing in t for every trajectory x(t) (the proof is the same as in Theorem 3). It remains to show that \(\psi \) is not identically zero. For an arbitrary \(x_0 \in \mathrm{int}\, K\), there is a constant \(c > 0\) such that inequality \(g(x) \le c\) implies \(x \, \le _K \, \frac{1}{2}\, x_0\). If \(\psi (x_0) = 0\), then there is \(n > 0\) such that \(r(x_0, n) < c\). Hence, there is a control function \(\bar{A} (\cdot )\) on the segment [0, n] and the corresponding trajectory \(\bar{x}\) with \(\bar{x}(0) = x_0\) and \(\bar{x}(n) \, \le _K \, \frac{1}{2}\, \bar{x}(0)\). If now \(A(\cdot )\) is the periodic extension of the control function \(\bar{A} (\cdot )\) to \({\mathbb {R}}_+\) with period n, then the corresponding trajectory \(x(\cdot )\) satisfies \(x(kn) \, \le _K \, 2^{-k}\, \bar{x}(0)\). Since the antinorm g is monotone on K (Lemma 8), we have \(g(x(kn)) \le 2^{-k} g(x_0)\, , \, k \in {\mathbb {N}}\). Moreover, g is positive, and hence it is equivalent to any norm on K. Therefore, \(\Vert x(kn)\Vert \le C\, 2^{-k}\, , \, k \in {\mathbb {N}}\), and hence \(\check{\sigma }\le - \frac{\ln 2}{n}\), which contradicts the assumption. Thus, \(\psi \) is an extremal antinorm, which concludes the proof of the first part.
Step 2. Now let us show that if all operators of \({\mathcal {A}}\) are K-irreducible, then there is a positive invariant antinorm. Take the extremal antinorm \(\psi \) on K constructed in the previous step and consider the function \(f(z) = \lim \limits _{t \rightarrow +\infty }r(z, t)\), where we now denote \(r(z, t) \, = \, \inf \, \{\psi (x(t))\, , \ A \in {\mathcal {U}}\, [0, t]\, , \, x(0) = z\}\). Since \(\psi (x(t))\, \) is non-decreasing in t, so is \(\, r(z, t)\). Hence, this limit exists for every \(z \in K\), maybe it becomes \(+\infty \). If \(f(x) = +\infty \) for some \(x \in K\), then f is equal to \(+\infty \) in the whole interior of K, since for every \(z \in \mathrm{int}(K)\) there is a constant c such that \(cz >_K x\).
Now we invoke Corollary 6: There exists a cone \(K'\) embedded in K such that every trajectory x(t) starting in \(K'\) remains in \(K'\). We see that if \(f(x) = +\infty \) for some \(x \in K\), then \(f(x_0) = +\infty \) for every point \(x_0 \in K'\setminus \{0\}\). Take an arbitrary \(x_0 \in K'\setminus \{0\}\). As we saw, \(f(x) = +\infty \) for some \(x \in K\), then \(f(x_0) = +\infty \). Since \(\psi \) is positive and continuous on \(K'\), there exists a constant \(C > 0\) such that for every \(x \in K'\) inequality \(\psi (x) \ge C\) implies \(x \, \ge _K \, 2\, x_0\). If \(f(x_0) = +\infty \), then there is \(q > 0\) such that \(r(x_0, q) > C\). Hence, \(x(q) \, \ge _K \, 2\, x(0)\) for every trajectory x(t) with \(x(0) = x_0\). Applying Corollary 4 and iterating k times, we get \(x(kq) \, \ge _K \, 2^{\, k}\, x_0\). Since \(\psi \) is monotone, it follows that \(\psi (x(kq)) \, \ge \, 2^{\, k}\psi (x_0)\). On the other hand, \(\psi \) is non-decreasing in t on every trajectory, consequently \(\psi (x(t)) \, \ge \, 2^{\bigl [\frac{t}{q} \bigr ]}\psi (x_0)\). Since \(\psi \) is equivalent to any norm on \(K'\), it follows that \(\check{\sigma }({\mathcal {A}}) \ge \frac{\ln 2}{n} > 0\). The contradiction shows that \(f(x)< +\infty \) for every \(x \in K\), i.e., f is an antinorm on K.
Consider now the optimization problem (28), where the maximum is replaced by minimum. It always has a solution \((\bar{x}, \bar{A})\, \in \, W_1^1\times L_1[0, \tau ]\) for which \(f(\bar{x} (\tau )) = F(z)\). Therefore, the set \(P_{\, \tau }\) of control functions \(A \in {\mathcal {U}}\, [0, \tau ]\) for which f(x(t)) is equal identically to f(z) on the segment \([0, \tau ]\) is non-empty. Since the sets \(\{P_{\, \tau }\}_{\, \tau \in {\mathbb {R}}_+}\) form an embedded system of non-empty compact sets, they have a common point \(\bar{A} (\cdot ) \in {\mathcal {U}}\, [0, +\infty )\), for which \(f(\bar{x}(t))=f(z), \,t\in [0, +\infty )\). Whence, f is an invariant antinorm. \(\square \)
Rights and permissions
About this article
Cite this article
Guglielmi, N., Laglia, L. & Protasov, V. Polytope Lyapunov Functions for Stable and for Stabilizable LSS. Found Comput Math 17, 567–623 (2017). https://doi.org/10.1007/s10208-015-9301-9
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10208-015-9301-9
Keywords
- Linear switching systems
- Lyapunov exponent
- Polytope
- Iterative method
- Cones
- Metzler matrices
- Joint spectral radius
- Lower spectral radius