Pith. sign in

REVIEW 2 major objections 5 minor 2 cited by

Feedback stabilization of switched systems under arbitrary switching: A convex characterization

T0 review · 2 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper proves that feedback stabilizability of a discrete-time switched linear system under arbitrary switching is equivalent to feasibility of a hierarchy of linear-matrix-inequality conditions built on feedback-tree graphs, for some…

desk verdict The sufficiency half and the feedback-tree framing are solid, but the necessity proof of the main characterization rests on a D_{w,i} existence assertion that is false for nilpotent closed-loop modes, so Theorem 2 is not yet established as stated. read the letter →

arxiv 2506.03759 v2 pith:T5BPTEA7 submitted 2025-06-04 math.OC cs.SYeess.SY

Thiago Alves Lima , Matteo Della Rossa , Antoine Girard This is my paper
classification math.OCcs.SYeess.SY MSC 93D1593D3093C5593C30
keywords switchedlinearsystemsarbitraryswitchingfeedbackstabilizationmatrixinequalitiesfeedback-treegraphsdifferenceinclusionsrobuststabilitymemory-dependentcontrollers
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that asking whether a discrete-time switched linear system can be stabilized by feedback, when the switching signal is an arbitrary disturbance, is exactly a convex feasibility question: the system is stabilizable precisely when a certain hierarchy of linear matrix inequalities, indexed by feedback-tree graphs, is feasible for some large enough graph order. This matters because earlier graph-based LMI conditions were only sufficient tests, and their necessity was an open question. The paper proves the equivalence for both mode-independent and mode-dependent feedback, and shows the stabilizing controllers can be taken piecewise linear, using min-of-quadratics Lyapunov functions, or linear with finite memory of the switching signal.

What carries the argument

The load-bearing object is the feedback-tree graph $\mathcal{F}^T(M)$: its nodes are all words over the mode alphabet of length at most $T$, including the empty word, and its edges go from a word $\hat{w}$ to $(\hat{w}, i)$ with label $i$ for words of length below $T$, and from length-$T$ words back to the root with every label. These graphs are complete and deterministic, which lets the existing graph-based Lyapunov sufficiency results apply directly. The associated LMIs assign positive-definite matrices $P_{\hat{w}}$ and gains $K_{\hat{w}}$ or $K_{\hat{w},i}$, and define the Lyapunov function $W(x)=\min_{\hat{w}}\sqrt{x^\top P_{\hat{w}}x}$. The finite-depth structure of the feedback tree is exactly what makes necessity provable: the robustness of stabilizability gives contraction after finitely many steps, and the columns of the constructed matrices $X_{\hat{w}}$ become the data that witness the LMIs.

What would settle it

Directly test the asserted existence step: pick any DFS system and stabilizing homogeneous feedback, choose $X_{\\hat{w}}$ by the recursive trajectory construction, and search numerically for $D_{\\hat{w},i}$ satisfying the row-space orthogonality, nonsingularity of $(A_i+B_iK_{\\hat{w},i})X_{\\hat{w}}+D_{\\hat{w},i}$, and the column bound $|d^j_{\\hat{w},i}|\\leq\\rho(x^j_{\\hat{w}})$; a single triple where every such $D$ violates the norm bound would break the necessity proof. Alternatively, run the feedback-tree LMIs on a candidate DFS system for $T=1,2,\\ldots$; any DFS system whose LMIs stay infeasible for every $T$ contradicts Theorem 2.

Watch

Extended reading notes

Core claim

The central claim is Theorem 2: the discrete-time switched system (3) is mode-independent feedback stabilizable (IFS) if and only if there is a large enough order $T$ for which the feedback-tree LMI conditions (13a)-(13b) are feasible, and mode-dependent feedback stabilizable (DFS) if and only if some $T$ makes (14a)-(14b) feasible. Sufficiency is obtained by applying known graph-Lyapunov propositions to the complete deterministic feedback-tree graphs, giving explicit piecewise-linear feedback laws and a continuous min-of-quadratics Lyapunov function. Necessity starts from the robust uniform exponential stability granted by the paper's Theorem 1, chooses $T$ so that the perturbed closed loop contracts after $T+1$ steps, and then builds nonsingular matrices $X_{\hat{w}}$ whose columns lie along perturbed trajectories; from these it constructs positive-definite $P_{\hat{w}}$ and gains $K_{\hat{w},i}$ satisfying the strict LMIs. The paper thus claims that no information is lost between stabilizability and the convex feasibility conditions.

Load-bearing premise

The proof of necessity needs, at every step, a perturbation matrix $D_{\\hat{w},i}$ whose row space is the orthogonal complement of a certain matrix's row space, that keeps a constructed matrix nonsingular, and whose columns stay within the robustness bound $\\rho$; the assertion that such a matrix can always be chosen is not proved, and the whole "only if" direction rides on it.

Editorial extensions

If this is right

  • If Theorem 2 is right, then any stabilizable system will eventually pass the feedback-tree LMI test, so searching over increasing graph order is a complete certification procedure for stabilizability.
  • The stabilizing feedback can always be chosen explicitly from an LMI solution as a piecewise-linear map $\Phi(x)=K_{\kappa(x)}x$, where $\kappa$ selects the minimizing quadratic, with $W$ as a continuous Lyapunov function.
  • The same LMI solution yields linear memory-dependent controllers using only the last $T$ values of the switching signal (T-SILm and T-SDLm), giving two different implementations with the same guaranteed decay rate.
  • If the LMIs are feasible at order $N$, they are also feasible at order $2N+1$ (Corollary 2), so the hierarchy has a doubling self-consistency property.
  • The $\\gamma$-scaled version of the conditions gives upper bounds on the optimal decay rates $\\gamma^I_*$ and $\\gamma^D_*$, and Theorem 2 implies these bounds are asymptotically exact as the graph order grows.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The theorem sharpens the old conjecture about De Bruijn graph conditions: since feedback-tree feasibility is necessary and De Bruijn conditions are sufficient, one could test whether De Bruijn feasibility implies feedback-tree feasibility, which would settle the conjecture directly.
  • The proof suggests that an explicit bound on $T$ in terms of the decay rate, the overshoot constant, the dimension, and the robustness margin might be extractable, turning the existential test into a finite one.
  • The unproved existence assertion for the matrices $D_{\\hat{w},i}$ is the single point where the necessity proof could fail; a small-dimensional numerical search for triples where every admissible $D$ violates the column norm bound would test that step.
  • If the upper bounds from the $\\gamma$-scaled LMIs are monotone in $T$, the numerical method gives a convergent sequence of computed decay-rate bounds that are certified to approach the true value.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper studies discrete-time switched linear systems with arbitrary switching treated as an exogenous disturbance, and it gives necessary and sufficient LMI conditions for mode-independent and mode-dependent feedback stabilizability. The main object is a hierarchy of 'feedback-tree graphs' of order T; Corollary 1 states that feasibility of the associated LMIs is sufficient for stabilizability, and Theorem 2 asserts that, for T large enough, the same LMIs are also necessary. The necessity proof assumes robust uniform exponential stability of a closed-loop inclusion, constructs nonsingular matrices X_w and gains K_w inductively, and then builds positive definite matrices P_w satisfying the strict LMIs. The paper also derives memory-based linear controllers from the same LMI solutions and reports two numerical examples.

Significance. If Theorem 2 is correct, the paper gives a convex characterization of feedback stabilizability under arbitrary switching: the problem reduces, at sufficiently high graph order, to feasibility of the feedback-tree LMIs. This is a substantial advance over the sufficient-only conditions in the authors' earlier De Bruijn graph approach, and the constructive necessity proof is valuable: it produces explicit P_w and K_w and clarifies the link between piecewise-linear controllers and memory-dependent linear controllers. The sufficiency direction is clean and follows from known graph-Lyapunov results, and the numerical examples illustrate the practical behavior of the hierarchy. The main caveat is that the necessity proof depends on two unproved existence/regularity steps, so the characterization is conditional until those steps are supplied.

major comments (2)
  1. [Theorem 2, proof after Eq. (27)] The assertion that 'it is always possible to choose a matrix D_{ŵ,i}' satisfying (28) and the norm bounds in item b) is load-bearing and is not proved. The orthogonality condition in (28) is used to eliminate the cross terms in the expansion of Q_{(ŵ,i)} in (31); the full-rank condition makes X_{(ŵ,i)} nonsingular, hence Q_{(ŵ,i)} positive definite and P_w well defined; and the norm bound is needed for the column inclusion (30), which in turn yields the contraction estimate (33). Without this existence step, the 'only if' direction of Theorem 2 does not go through. The claim is plausible -- one can choose the rows of D_{ŵ,i} to be a small basis of ker((A_i+B_iK_{ŵ,i})X_ŵ) and scale them so that the column norms are below the corresponding ρ-values -- but the scaling must also preserve invertibility of the sum, and this has to be demonstrated. I recommend adding a lemma, with a constructive proof, immediately before Eq. (28).
  2. [Appendix A, Lemma 2] The proof of Lemma 2 asserts that the infimum in (44) is attained because V is radially unbounded and continuous, and that a minimizer Φ(x) can be chosen to be homogeneous of degree 1. Neither step is immediate. If the matrices B_i have a common kernel, the map u ↦ max_i V(A_i x+B_i u) is constant along that kernel, so coercivity holds only on a quotient and the attainment argument needs a compactness argument on that quotient. More importantly, the existence of a homogeneous selector with the regularity needed for Lemma 1 is not established: Lemma 1's second point requires the closed-loop map F^Φ to be nonempty-valued, closed, and locally bounded, and these hypotheses are not verified for an arbitrary homogeneous selector. Since Theorem 1 is used in the necessity part of Theorem 2, this gap should be closed either by proving existence of a suitable selector (for instance by a measurable-selection or tie-breaking construction) or by proving RUES directly from the continuous Lyapunov inequality without relying on the closedness assumption.
minor comments (5)
  1. [Title page] The first author's name is spelled 'Thaigo Alves Lima'; this appears to be a typo for 'Thiago'.
  2. [Corollary 1] The notation K_ŵ = K_ŵP_ŵ overloads the symbol K_ŵ; the gain on the left is an m×n matrix while the K_ŵ on the right is the LMI variable. The same overloading appears in the mode-dependent case with K_{ŵ,i}.
  3. [Proposition 3, proof] In the last paragraph of the proof, the expression K(cut_{T+1}(σ,k−1), σ(k)) should be K(cut_{T+1}(σ,k), σ(k)) to match the definition in (37).
  4. [Section 5.2] The sentence 'the conditions of Proposition 1 proposed in this paper' should refer to Corollary 1 (or Theorem 2), since Proposition 1 is recalled from [9] and is not proposed in this paper.
  5. [Lemma 1] The proof of item 2 is said to be 'straightforward' and is omitted; a citation to the specific theorem in [21] or a short proof sketch would improve verifiability, especially because Lemma 1 is invoked in the delicate selector argument of Lemma 2.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Theorem 2's necessity is a genuine construction from RUES, and the cited [9] sufficiency results are independent support; the flagged proof gap after eq. (28) is not circular.

full rationale

Systematic walk of the derivation chain finds no circular step. Sufficiency in Theorem 2 is obtained through Corollary 1 as an application of Propositions 1 and 2 from [9]; although [9] shares authors with the present paper, those propositions are published sufficient conditions with their own proofs, are parameter-free, and do not assume the target equivalence, so they are independent evidence rather than a load-bearing self-citation. The necessity direction is the substantive new contribution: starting from DFS and Theorem 1, the proof assumes only RUES and explicitly constructs X_w, K_w,i, D_w,i, Q_w, and P_w; the LMI conditions (14) are the conclusion of the construction, not a premise or a fitted quantity renamed as a prediction. No uniqueness theorem or ansatz is imported from the authors to force the choice. The memory-controller results in Proposition 3 are direct consequences of the LMI solutions and introduce no circularity. The passage that should be flagged is the assertion after equation (27) that 'it is always possible to choose a matrix D_w,i' satisfying (28) and the norm bound |d^j_w,i| ≤ ρ(x^j_w). This assertion is load-bearing for the recursive construction and is not proved in the manuscript; however, it is a completeness/correctness gap, not a circularity, because it does not restate the target theorem and does not presuppose LMI feasibility. The remaining external reliance on [9, 15, 21] is real evidence, so the paper merits a score of 0 for circularity.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The main theorem introduces no fitted constants or ad hoc numerical parameters. It rests on standard converse Lyapunov results, on the authors' earlier graph-based sufficiency results, and on two unproved technical assertions in the proof (D_w,i existence and homogeneous selector/dynamic programming details). These are the main correctness burden and are likely repairable.

assumptions (6)
  • domain assumption Difference inclusion representation x+ in {A_i x+B_i u | i in <M>} is an exact representation of arbitrary switching for stabilization purposes.
    Used from Section 3 onward to turn arbitrary switching into a set-valued system; exact for worst-case stabilization because the controller does not know future modes.
  • standard math Converse Lyapunov theorems for difference inclusions (Lemma 1) hold as stated.
    Imported from [15,21]: UES iff exponential Lyapunov function exists; continuous Lyapunov function implies RUES. The paper does not reprove them.
  • domain assumption Propositions 1 and 2 of [9]: complete/deterministic graph inequalities imply exponential stabilization by piecewise-linear feedback.
    Used as a black box for the sufficiency direction of Corollary 1; [9] is a published result by two of the authors.
  • ad hoc to paper The dynamic programming principle (44) holds and the infimum over controllers is attained.
    Invoked in Lemma 2 to construct the value function and the homogeneous feedback; the paper sketches this with a citation to [3] but does not justify the interchange of sup over switching signals and inf over controllers.
  • ad hoc to paper Existence of D_w,i matrices satisfying rank, orthogonality, and norm constraints.
    Asserted without proof in the necessity proof of Theorem 2; needed to maintain nonsingularity of all X_w.
  • ad hoc to paper A homogeneous selector from the argmin sets in Lemma 2 exists.
    Homogeneity of the value function is shown, but the existence of a homogeneous measurable selector is not demonstrated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Feedback stabilization of switched systems under arbitrary switching: A convex characterization." pith.science (2026). https://pith.science/paper/T5BPTEA7

@misc{pith2026250603759,
  author       = {Pith},
  title        = {Pith review of: Feedback stabilization of switched systems under arbitrary switching: A convex characterization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T5BPTEA7}},
  note         = {Machine review of arXiv:2506.03759}
}
read the original abstract

In this paper, we study stabilizability of discrete-time switched linear systems where the switching signal is considered as an arbitrary external input (and not a control variable). We characterize feedback stabilization via a hierarchy of necessary and sufficient linear matrix inequalities (LMIs) conditions based on novel graph structures. We analyze both the cases in which the controller has (or has not) access to the current switching mode, the so-called mode-dependent and mode-independent settings, providing specular results. Moreover, our approach provides explicit piecewise-linear and memory-dependent linear controllers, highlighting the connections with existing stabilization approaches. The effectiveness of the proposed technique is finally illustrated with the help of some numerical examples.

Figures

Figures reproduced from arXiv: 2506.03759 by the authors.

Figure 1
Figure 1. The graphs F 1 (2), F 2 (2), and F 3 (2), i.e., the feedback-tree graphs of order 1, 2, and 3 on ⟨M⟩ = {1, 2}. The edges labeled by multiple labels can be considered as a graphical shortcut for multiple edges, each one with a single label. The proofs of Propositions 1 and 2 can be found in [9]. We note here that the robustness of the stable closed loop, already ensured by Theorem 1, can also be corroborated via Lemm… view at source ↗
Figure 2
Figure 2. The graphs G 2 DB on the alphabet {1, 2}. the proof of [15, Theorem 7] implies that system (3) equipped with the memory-dependent linear controller in (37) is uniformly exponentially stable (in the sense of Definition 7). Moreover, since K(k, σ[0,k]) = K(cutT +1(σ,k−1), σ(k)) and the function cutT +1 defined in (35) only depends, at most, on [σ][k+1−T −1,k] = [σ][k−T ,k] we have that the system is stabilizable mode … view at source ↗
Figure 3
Figure 3. Results for Example 5.1. the function κ : R n → SM,T (introduced in Corollary 1), which selects, for each time x(k), the control that minimizes the current “energy” of the system measured by the Lyapunov function W. Such minimization allows for “jumps” among the nodes of the FTG, therefore leading to different values for W and V . From Fig. 3a, it is interesting to note that although such minimization happens at eac… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Schematic collection of the main results in the mode-independent case. The mode-dependent case [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Feedback Stabilization of Switched Systems: Memory is not needed

    math.OC 2026-05 unverdicted novelty 7.0 of 10

    Proves that stabilizing switched linear systems with full-history controllers implies existence of memoryless homogeneous degree one controllers depending only on current state (and mode if applicable).

  2. Robust Optimal Control of Arbitrarily Switched Systems: A Path-Complete Framework

    math.OC 2026-07 conditional novelty 6.0 of 10

    A path-complete graph framework jointly synthesizes feedback policies and certified upper bounds on the infinite-horizon value function for arbitrarily switched systems, with SDP/alternating-optimization formulations ...

Reference graph

Works this paper leans on

32 extracted references · 32 canonical work pages · cited by 2 Pith papers

  1. [9]

    Della Rossa, T

    M. Della Rossa, T. Alves Lima, M. Jungers, and R. M. Jungers. Graph-based conditions for feedback stabilization of switched and LPV systems. Automatica, 160:111427, 2024

    work page 2024

  2. [1]

    A. A. Ahmadi, R. M. Jungers, P. A. Parrilo, and M. Roozbehani. Joint spectral radius and path-complete graph Lyapunov functions. SIAM Journal on Control and Optimization , 52:687–717, 2014

    work page 2014

  3. [2]

    E. D. Andersen and K. D. Andersen. The MOSEK interior point optimizer for linear programming: An implementation of the homogeneous algorithm. In H. Frenk, K. Roos, T. Terlaky, and S. Zhang, editors, High Performance Optimization , volume 33 of Applied Optimization, pages 197–232. Springer US, 2000. http://www.mosek.com

    work page 2000

  4. [3]

    Bertsekas

    D. Bertsekas. Dynamic programming and optimal control: Volume I , volume 4. Athena scientific, 2012

    work page 2012

  5. [4]

    Blanchini and S

    F. Blanchini and S. Miani. Stabilization of LPV systems: State feedback, state estimation, and duality. SIAM Journal on Control and Optimization , 42(1):76–97, 2003

    work page 2003

  6. [5]

    Blanchini and S

    F. Blanchini and S. Miani. Set-theoretic methods in control, volume 78. Springer, 2008

    work page 2008

  7. [6]

    Blanchini, S

    F. Blanchini, S. Miani, and C. Savorgnan. Stability results for linear parameter varying and switching systems. Automatica, 43(10):1817–1823, 2007

    work page 2007

  8. [7]

    Daafouz and J

    J. Daafouz and J. Bernussou. Parameter dependent lyapunov functions for discrete time systems with time varying parametric uncertainties. Systems & Control Letters , 43(5):355–359, 2001

    work page 2001

Show all 32 references
  1. [8]

    N. G. de Bruijn. A combinatorial problem. Proceedings of the Section of Sciences of the Koninklijke Nederlandse Akademie van Wetenschappen te Amsterdam , 49(7):758–764, 1946

    1946

  2. [10]

    Della Rossa and R.M

    M. Della Rossa and R.M. Jungers. Multiple Lyapunov functions and memory: A symbolic dynamics approach to systems and control. SIAM Journal on Control and Optimization , 62(5):2695–2722, 2024

    2024

  3. [11]

    C. P. Dettmann, R. M. Jungers, and P. Mason. Lower bounds and dense discontinuity phenomena for the stabilizability radius of linear switched systems. Systems & Control Letters , 142:104737, 2020

    2020

  4. [12]

    Essick, J.-W

    R. Essick, J.-W. Lee, and G. E. Dullerud. Control of linear switched systems with receding horizon modal information. IEEE Transactions on Automatic Control , 59(9):2340–2352, 2014

    2014

  5. [13]

    Fiacchini, A

    M. Fiacchini, A. Girard, and M. Jungers. On the stabilizability of discrete-time switched linear systems: Novel conditions and comparisons. IEEE Transactions on Automatic Control , 61(5):1181–1193, 2016

    2016

  6. [14]

    Fiacchini and M

    M. Fiacchini and M. Jungers. Necessary and sufficient condition for stabilizability of discrete-time linear switched systems: A set-theory approach. Automatica, 50(1):75–83, 2014

    2014

  7. [15]

    Geiselhart, R.H

    R. Geiselhart, R.H. Gielen, M. Lazar, and F. R. Wirth. An alternative converse Lyapunov theorem for discrete-time systems. Systems & Control Letters , 70:49–59, 2014

    2014

  8. [16]

    Goebel, T

    R. Goebel, T. Hu, and A. R. Teel. Dual Matrix Inequalities in Stability and Performance Analysis of Linear Differential/Difference Inclusions, pages 103–122. Birkh¨ auser Boston, 2006

    2006

  9. [17]

    J. Hu, J. Shen, and D. Lee. Resilient stabilization of switched linear control systems against adversarial switching. IEEE Transactions on Automatic Control , 62(8):3820–3834, 2017

    2017

  10. [18]

    J. Hu, J. Shen, and D. Lee. Optimal stabilizing rates of switched linear control systems under arbitrary known switchings. Automatica, 159:111331, 2024

    2024

  11. [19]

    Hu and F

    T. Hu and F. Blanchini. Non-conservative matrix inequality conditions for stability/stabilizability of linear differential inclusions. Automatica, 46(1):190–196, 2010. 20

    2010

  12. [20]

    Jungers and P

    R.M. Jungers and P. Mason. On feedback stabilization of linear switched systems via switching signal control. SIAM Journal on Control and Optimization , 55(2):1179–1198, 2017

    2017

  13. [21]

    Kellett and A.R

    C.M. Kellett and A.R. Teel. On the robustness of KL-stability for difference inclusions: Smooth discrete- time Lyapunov functions. SIAM Journal on Control and Optimization , 44(3):777–800, 2005

    2005

  14. [22]

    Lee and G.E

    J.-W. Lee and G.E. Dullerud. Uniform stabilization of discrete-time switched and Markovian jump linear systems. Automatica, 42(2):205–218, 2006

    2006

  15. [23]

    Lee and P.P

    J.-W. Lee and P.P. Khargonekar. Detectability and stabilizability of discrete-time switched linear sys- tems. IEEE Transactions on Automatic Control , 54(3):424–437, 2009

    2009

  16. [24]

    Liberzon

    D. Liberzon. Switching in Systems and Control . Systems & Control: Foundations & Applications. Birkh¨ auser, 2003

    2003

  17. [25]

    Lin and P

    H. Lin and P. J. Antsaklis. Hybrid state feedback stabilization with ℓ2 performance for discrete-time switched linear systems. International Journal of Control , 81(7):1114–1124, 2008

    2008

  18. [26]

    Lin and P.J

    H. Lin and P.J. Antsaklis. Stability and stabilizability of switched linear systems: A survey of recent results. IEEE Transactions on Automatic Control , 54(2):308–322, 2009

    2009

  19. [27]

    L¨ ofberg

    J. L¨ ofberg. YALMIP: A toolbox for modeling and optimization in MATLAB. InProceedings of the 2004 IEEE International Symposium on Computer Aided Control Systems Design , pages 284–289, Taipei, Taiwan, September 2004. http://yalmip.github.io

    2004

  20. [28]

    Philippe, N

    M. Philippe, N. Athanasopoulos, D. Angeli, and R.M. Jungers. On path-complete Lyapunov functions: Geometry and comparison. IEEE Transactions on Automatic Control , 64(5):1947–1957, 2019

    1947

  21. [29]

    Philippe, R

    M. Philippe, R. Essick, G.E. Dullerud, and R.M. Jungers. Stability of discrete-time switching systems with constrained switching sequences. Automatica, 72:242–250, 2016

    2016

  22. [30]

    Polyakov

    A. Polyakov. Generalized Homogeneity in Systems and Control . Communications and Control Engi- neering. Springer International Publishing, 2020

    2020

  23. [31]

    Rockafellar and R.J.-B

    R.T. Rockafellar and R.J.-B. Wets. Variational Analysis, volume 317 of Gundlehren der mathematischen Wissenchaften. Springer-Verlag, Berlin, 3rd printing, 2009 edition, 1998

    2009

  24. [32]

    if part

    W. Zhang, A. Abate, J. Hu, and M. P. Vitus. Exponential stabilization of discrete-time switched linear systems. Automatica, 45(11):2526–2536, 2009. A Proof of Theorem 1 In this section, we provide the formal proof of Theorem 1. The “if part ” is a straightforward consequence o...

    2009

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.