Decoupling Periodic Systems: An Algebraic Approach

Vladim\'ir Ku\v{c}era

REVIEW 3 major objections 7 minor 1 references

Decoupling Periodic Systems: An Algebraic Approach

T0 review · 3 major / 7 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 Theorem 2 gives the first necessary and sufficient conditions for row-by-row decoupling of nonsquare periodic systems with more inputs than outputs via nonregular periodic state feedback.

desk verdict A plausible first solution to nonsquare periodic decoupling, but the central sufficiency proof leans on an unproved block-diagonal realizability claim. read the letter →

arxiv 2505.23616 v2 pith:7UHOUHTA submitted 2025-05-29 eess.SY cs.SYmath.RA

Vladimír Kučera This is my paper

classification eess.SYcs.SYmath.RA MSC 93C5593C0593B5293D15

keywords row-by-rowdecouplingperiodicsystemsnonregularstatefeedbackcyclicrepresentationHermitenormalforminteractorstability

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

The paper tackles the row-by-row decoupling problem for discrete-time linear systems with periodically time-varying coefficients: find periodic state feedback so that each output is controlled by exactly one input while stability is preserved. Prior work covered square systems by dynamic compensation and block decoupling by regular feedback; the case of more inputs than outputs, which requires nonregular feedback, was open. The paper proves, as Theorem 2, necessary and sufficient conditions for this case, expressed as list conditions on the cyclic time-invariant representation of the periodic system. The conditions are stated in terms of decoupling orders, invariant factors, and the lengths of unobservable pure-delay chains, and the proof constructs the nonregular periodic feedback explicitly. A sympathetic reader would care because it closes a gap in the decoupling theory of periodic systems and gives a constructive algorithm, not merely an existence test.

What carries the argument

The machinery is the cyclic representation of a periodic system: an equivalent time-invariant system of dimension nT whose transfer function W_tau(z) is a pT by mT matrix with a block structure inherited from the periodic data. Work is done in the ring S of proper and stable rational functions, whose units are the unimodular matrices; the cyclic Hermite normal form over S is the canonical form under block-diagonal unimodular column operations. The decisive objects are the decoupling standard form, obtained by reducing the cyclic transfer function to Hermite form and isolating the maximally unobservable subsystem; the cyclic interactor, the inverse of the matrix of leading decoupling entries, which carries the infinite and unstable zeros; and the lists delta_i, phi_i, sigma_i, epsilon_i, eta_i that record decoupling orders, interactor orders, reachability indices of the unobservable chains, required order increases, and chain connections. The key restriction is that all transformations must be block diagonal at infinity so they lift to genuine periodic state feedback.

What would settle it

Check whether there exists a stable periodic system with m > p whose cyclic transfer function has a decoupling compensator that is only realizable by a non-block-diagonal unimodular transformation, while the invariant lists (16)-(18) of Theorem 2 fail; this would directly refute necessity. A more computational falsifier is to run the construction of steps (8)-(12) on a system where the lists satisfy the theorem but no periodic choices of eta_i and epsilon_i terminate the search, which would expose a gap in sufficiency.

Watch

Extended reading notes

Core claim

The central claim is that the Decoupling Problem for a stable periodic system with more inputs than outputs is solvable by periodic state feedback if and only if there exist lists of nonnegative integers (16), each entry a multiple of the period T, together with sublists (17) and (18), such that the decoupling order lists, the interactor pole lists, and the unobservable chain lengths satisfy the inequalities and index-set equalities of Theorem 2. This is the first necessary and sufficient condition for stability-preserving row-by-row decoupling of nonsquare periodic systems via nonregular state feedback. For square systems, the paper shows in Theorem 1 that decoupling is possible exactly when all blocks of the cyclic Hermite normal form of the cyclic transfer function are diagonal, which is realized by regular feedback. The construction shows how extra input terminals are connected to the unobservable pure-delay chains, raising decoupling orders and supplying the missing unstable invariant zeros, while the 'multiple of T' requirement preserves periodicity.

Load-bearing premise

The proof requires that every decoupling compensator for the cyclic time-invariant system can be taken to act in a block-diagonal way at infinity, so that it translates back into a periodic state feedback; the paper asserts this follows from the block structure but does not give a complete proof.

Editorial extensions

If this is right

A constructive algorithm now exists that either produces the periodic nonregular state feedback decoupling a given stable nonsquare periodic system, or certifies that none exists.
The square-system case is recovered as a special case: regular feedback suffices, and the condition is that the cyclic Hermite normal form of the cyclic transfer function be block-diagonal.
The decoupling compensator built by the proof makes the monodromy matrix of the closed-loop system have all eigenvalues equal to zero; if the system is reachable, the core spectrum can then be assigned at will.
The necessary and sufficient conditions are invariant under state feedback and input and state coordinate transformations, so they describe an intrinsic property of the periodic system rather than of a particular representation.
Systems with more outputs than inputs remain block-decouplable only, consistent with the time-invariant theory; the theorem clarifies exactly when the extra inputs allow full diagonal decoupling.

Reading between the lines

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

Editorial extension: The proof's block-diagonality-at-infinity assumption is the place a counterexample would most likely hide; a non-block-diagonal compensator that decouples the cyclic system while evading the list conditions would refute the theorem as stated.
Editorial extension: Since the cyclic representation does not depend on the initial time instant's zero structure, the same conditions should characterize decoupling for every choice of the initial sampling time, which the paper indicates but does not spell out as an invariance theorem.
Editorial extension: A natural testable extension is to continuous-time periodic systems, where the role of the ring S of proper stable rational functions would be played by a suitable ring of exponentially stable transfer functions and the pure-delay chains by chains of integrators.
Editorial extension: The theorem presupposes the given system is already stable; applying it to stabilizable systems requires a preliminary stabilizing feedback, and the paper notes this may destroy stabilizability, so a combined stabilization-and-decoupling condition remains an open question.

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.

Desk Editor's Note

read the letter

The genuinely new thing here is the nonsquare, more-inputs-than-outputs case of row-by-row decoupling for periodic systems, solved by nonregular periodic state feedback. The claim of first solution looks correct. The square case is a cleaner alternative to earlier results, and Theorem 1 is proved carefully. All three examples are fully worked, which helps a lot.

The soft spot is Theorem 2, the main result. Its proof is a sketch: it calls the transfer from [6] a 'slight refinement' and pins everything on the sentence that the unimodular transformations 'must also be block diagonal at infinity.' That is exactly the step that makes the time-invariant compensator realizable as periodic state feedback. Necessity is fine—if periodic state feedback decouples, the cyclic compensator has the right block structure. But sufficiency requires showing that the constructed U(z) from the time-invariant algorithm satisfies U^{-1}(z)-U^{-1}(∞)=K S(z) with block-diagonal constant K. The paper does not prove that; step (12) just asserts L, V(∞) and Z(∞) can be made block-diagonal. The T-multiplicity condition in (16) presumably is what guarantees this, but the text never shows the mechanics. The examples are consistent, but they are not a substitute for the missing argument.

The reader's stress-test note is on target. I don't think this is a fatal flaw—the result is likely true and the author is working in his own well-established framework—but the paper as written is incomplete at the pivot point. A referee should request an expanded proof of Theorem 2, or an explicit statement of the version of [6, Theorem 1] used plus a lemma proving the block-diagonal realizability.

The citation pattern is honest: [6] is his own prior theorem, and using it to transfer to periodic systems is legitimate, not circular. There is no code or formal proof artifact, which is normal for this kind of paper.

Who is this for? People working on periodic systems and algebraic decoupling. It is a solid within-subfield contribution, not a paradigm shift. I would send it to review rather than desk reject; the gap is specific and likely fixable, and the claimed first solution deserves referee time.

Referee Report

3 major / 7 minor

Summary. The paper studies row-by-row (diagonal) decoupling of discrete-time linear periodic systems by periodic state feedback, with particular attention to nonsquare systems having more inputs than outputs, where nonregular feedback is necessary. The approach is purely algebraic: it embeds the periodic system into its cyclic time-invariant representation and works over the ring S of proper stable rational functions. Theorem 1 gives a necessary and sufficient condition for square stable systems in terms of the cyclic Hermite normal form having diagonal blocks. Theorem 2, the main result, extends this to nonsquare systems with m > p, stating a necessary and sufficient condition expressed through lists of nonnegative integers (16)-(18) whose elements are multiples of T, together with inequalities involving the decoupling orders and invariant factors. The proof of Theorem 2 is presented as a 'slight refinement' of the author's earlier time-invariant result [6, Theorem 1], with the construction summarized in algorithmic steps (8)-(13). Three worked examples illustrate the square case and two nonsquare cases.

Significance. If Theorem 2 is correct, it provides the first necessary and sufficient condition for row-by-row decoupling of stable discrete-time periodic systems with more inputs than outputs via nonregular periodic state feedback. The algebraic reduction to a time-invariant cyclic representation is a natural and potentially powerful framework, and the paper's worked examples give concrete evidence for the construction. The paper also provides a self-contained statement and proof for the square case in Theorem 1. However, the proof of Theorem 2 is only a sketch: the central transfer from the time-invariant theorem to the periodic setting rests on an unproved block-diagonal realizability assertion, and the derivation of the periodic list conditions is not given. Thus the paper's main claim is currently not fully supported.

major comments (3)

[Section VII, proof of Theorem 2 and step (12)] The sufficiency of Theorem 2 depends on the assertion that the unimodular compensator obtained from [6, Theorem 1] can be chosen to be block diagonal at infinity. The proof contains the sentence 'Consequently, the corresponding unimodular transformations must also be block diagonal at infinity,' and step (12) states that 'elementary operations can also be applied to make V(∞) and Z(∞) block-diagonal as well.' This is the load-bearing loss-of-generality claim: periodic state feedback imposes the constraint Uτ^{-1}(z) - Uτ^{-1}(∞) = K Sτ(z) with K block diagonal, exactly as in the proof of Theorem 1. The paper does not prove that the construction in steps (8)-(13) can always satisfy this equation while preserving the matrix identity (23). If every decoupling compensator for some system satisfying (16)-(18) were forced to be non-block-diagonal at infinity, the stated conditions would be insufficient. A complete proof of this realizability claim is required.
[Section VII, proof of Theorem 2] The statement 'The proof involves a slight refinement of the result [6, Theorem 1] applicable to time-invariant systems' is not a proof. The manuscript does not derive the periodic list conditions (16)-(18) from the time-invariant conditions of [6], does not justify why each element of the lists in (16) must be a multiple of T, and does not explain how the inequalities involving (17)-(20) follow from the cyclic structure. The paragraph describing the 'multi-diagonal' structure of the cyclic Hermite normal form and the interconnection of delay chains is too vague to serve as a derivation. A self-contained proof of Theorem 2 is needed, not merely a reference to the author's earlier theorem.
[Section V, problem formulation] The paper asserts that 'We cannot destroy the existence of a decoupling state feedback when applying stabilizing state feedback' without proof or reference. This is a nontrivial invariance property. The main theorems are stated for stable systems, so this assertion is not needed for Theorem 2 as stated, but the introduction promises an 'unrestricted' decoupling solution. The authors should either prove this invariance or clearly limit the scope of the paper to systems that are already stable.

minor comments (7)

[Header] The running header 'First Author et al.: Title' appears throughout the manuscript and should be replaced with the actual title and author information.
[Section VII, Theorem 2] The conditions following 'such that the following conditions are satisfied for each i = 1, 2, …, T:' are not displayed in the submitted text; the theorem statement is incomplete as printed and should show the explicit inequalities.
[Section IV] The definition of the normal form for nonassociates is missing its displayed expression; the text should show the form e/z^k or the equivalent.
[Section III] The sentence 'The eigenvalues of are determined by the Tth root...' has a missing matrix symbol, presumably Aτ, and should be corrected.
[Section VII, step (12)] The phrase 'Z(s)-matrix' should be 'Z(z)-matrix' for consistency with the variable z used throughout the paper.
[Section VII, step (13)] The sentence 'The monodromy matrix of the decoupled system has all eigenvalues equal to zero; if the system is reachable, we can change its core spectrum at will' is internally inconsistent: if all eigenvalues are zero, there is no nonzero core spectrum left to change. Please clarify which part of the monodromy has zero eigenvalues and how the core spectrum is assigned.
[Example 2] The sentence 'The system (25) is in the decoupling standard form with The unobservable chain is free.' contains a gap; the missing clause should be supplied.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Theorem 2 is a legitimate reduction to the author's prior time-invariant theorem [6], and the flagged block-diagonal realizability step is an unproved assertion, not a self-referential one.

full rationale

I examined the claimed derivation chain for equivalence between inputs and conclusions. Theorem 2's proof explicitly builds on [6, Thm. 1], a published time-invariant decoupling theorem. The periodic target is not an assumption of [6]; the cyclic representation is a standard isomorphism, so this is an independent, non-circular dependency. The decoupling standard form, cyclic interactor, and lists (11)-(19) are computed from the system's cyclic transfer function and invariant structure; the theorem's conditions (16)-(18) are criteria, not fitted parameters later renamed as predictions. The only load-bearing step that could be questioned is the assertion in Sec. VII step (12) that 'elementary operations can also be applied to make V(∞) and Z(∞) block-diagonal as well.' That is a loss-of-generality claim on which sufficiency rests, but it is a correctness gap or unproven lemma, not circularity: the claim does not presuppose the conclusion or define the target in terms of itself. Likewise, the reliance on [6] is self-citation, but [6] is a prior independent peer-reviewed result; no uniqueness theorem or ansatz is smuggled in. Hence no circular step, score 0.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The central claim rests on three legs: the cyclic representation equivalence from [10], the Hermite normal form theory over S from [11], and the time-invariant decoupling theorem [6, Theorem 1]. The first two are standard published background. The third is the same author's prior result, used as a black box; this is not circular but shifts much of the proof burden outside the paper. The paper also assumes, without proof, that pre-stabilization cannot destroy decouplability.

assumptions (4)

domain assumption The cyclic representation faithfully embeds the periodic system and preserves stability and the zero structure.
Invoked in Section III to justify reducing the periodic problem to a time-invariant one; based on Bittanti and Colaneri [10].
domain assumption The Euclidean domain S of proper stable rational functions and its Hermite normal form theory apply to cyclic transfer functions while preserving the block structure through block-diagonal unimodular transformations.
Used in Sections IV and VI; relies on [11] for Hermite normal form over S and on the author's [6] for the decoupling theorem.
domain assumption [6, Theorem 1] gives necessary and sufficient conditions for decoupling of time-invariant systems with more inputs than outputs.
Theorem 2's proof is explicitly a 'slight refinement' of this theorem; the current paper does not reprove it.
ad hoc to paper Preliminary stabilizing state feedback does not destroy the existence of a decoupling feedback, and nonregular decoupling feedback may destroy stabilizability.
Section V asserts these properties without proof or reference; they underpin the restriction of the theorems to stable systems.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Decoupling Periodic Systems: An Algebraic Approach." pith.science (2026). https://pith.science/paper/7UHOUHTA

@misc{pith2026250523616,
  author       = {Pith},
  title        = {Pith review of: Decoupling Periodic Systems: An Algebraic Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7UHOUHTA}},
  note         = {Machine review of arXiv:2505.23616}
}

read the original abstract

This paper addresses the problem of row-by-row (or diagonal) decoupling of discrete-time linear multi-input multi-output systems with periodic time-varying coefficients using periodic state feedback. Previous solutions have tackled row-by-row decoupling using dynamic compensation for square systems and block-decoupling through regular state feedback for nonsquare systems with more outputs than inputs. While it appears likely that a row-by-row state feedback solution for square systems can be deduced from these findings, a direct argument seems more appropriate here as it presents a natural extension for decoupling nonsquare systems with more inputs than outputs. This extension, which necessitates nonregular state feedback, has yet to be explored for periodic systems. Our approach is purely algebraic, based on a time-invariant representation of the periodic system.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

1 extracted references · 1 canonical work pages

[2006]

η1=(1), η2=(1); η1∗=(0,0), η2∗=(1,0)

W(z)=1000000z−20z−1000001000z−10010⎡⎣⎢⎢⎢⎢⎢⎤⎦⎥⎥⎥⎥⎥Φ(z)=10000z2−z000100001⎡⎣⎢⎢⎢⎢⎢⎤⎦⎥⎥⎥⎥⎥δ1=(0,2),δ2=(0,0);ϕ1=(0,2),ϕ2=(1,0)d1=(1,1),d2=(1,1);f1=(1,1),f2=(1,1).σ1=(1), σ2=(2).σ1f=(1), σ2f=(1).ε1=(0,0), ε2=(2,0). η1=(1), η2=(1); η1∗=(0,0), η2∗=(1,0). Z1(z)=100001−z−1000z−200001⎡⎣⎢⎢⎢⎢⎢⎤⎦⎥⎥⎥⎥⎥ω1=(0,0),ω2=(0,2).V22(z)=1−z−1−z−10⎡⎣⎢⎢⎤⎦⎥⎥Z1(z),η1=(1), η2=(1).V21(z...

work page arXiv 2001

Pith tools

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

[1] [2006]

η1=(1), η2=(1); η1∗=(0,0), η2∗=(1,0)

W(z)=1000000z−20z−1000001000z−10010⎡⎣⎢⎢⎢⎢⎢⎤⎦⎥⎥⎥⎥⎥Φ(z)=10000z2−z000100001⎡⎣⎢⎢⎢⎢⎢⎤⎦⎥⎥⎥⎥⎥δ1=(0,2),δ2=(0,0);ϕ1=(0,2),ϕ2=(1,0)d1=(1,1),d2=(1,1);f1=(1,1),f2=(1,1).σ1=(1), σ2=(2).σ1f=(1), σ2f=(1).ε1=(0,0), ε2=(2,0). η1=(1), η2=(1); η1∗=(0,0), η2∗=(1,0). Z1(z)=100001−z−1000z−200001⎡⎣⎢⎢⎢⎢⎢⎤⎦⎥⎥⎥⎥⎥ω1=(0,0),ω2=(0,2).V22(z)=1−z−1−z−10⎡⎣⎢⎢⎤⎦⎥⎥Z1(z),η1=(1), η2=(1).V21(z...

work page arXiv 2001