pith. sign in

arxiv: 2502.13523 · v1 · submitted 2025-02-19 · 🧮 math.OC

An application of the mean motion problem to time-optimal control

Omri Dalin , Alexander Ovseevich , Michael Margaliot This is my paper

Pith reviewed 2026-05-23 02:50 UTC · model grok-4.3

classification 🧮 math.OC
keywords time-optimal controlswitching pointsmean motion problemlinear systemsimaginary eigenvaluesbang-bang control
0
0 comments X

The pith

For linear systems with purely imaginary eigenvalues, the number of switches in time-optimal controls grows linearly with the time interval length in generic cases.

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

The paper studies time-optimal controls for controllable linear systems driven by a single scalar input over long time intervals. It is known that when all eigenvalues of the system matrix are real the number of switching points stays bounded no matter how long the interval becomes. The authors consider the complementary case of purely imaginary eigenvalues and prove that, under a generic condition, the number of switches is bounded from below by a positive multiple of the interval length. The argument works by identifying the switching function with an instance of the classical mean-motion problem whose solution goes back to Weyl. A reader would care because the result shows that oscillatory linear dynamics force optimal controls to become steadily more complicated as the horizon lengthens.

Core claim

For a controllable linear system with scalar control whose dynamics matrix has purely imaginary eigenvalues, any time-optimal control on a time interval of length T has, in the generic case, a number of switching points bounded from below by a positive constant times T.

What carries the argument

The switching function whose zeros mark the control switches, identified with the mean-motion problem of counting the average number of revolutions of a linear combination of rotating vectors.

If this is right

  • Longer time horizons force proportionally more switches in the optimal control whenever the dynamics are oscillatory.
  • The linear lower bound holds uniformly for almost all initial states under the stated genericity condition.
  • This lower bound stands in direct contrast to the known upper bound that is independent of horizon length when eigenvalues are real.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Numerical synthesis of optimal controls for resonant mechanical or electrical systems will need to accommodate an increasing number of switches as the planning horizon grows.
  • The result may extend to nearby problems such as minimum-time steering of harmonic oscillators or undamped linear networks.
  • Adding small damping or nonlinear terms could be tested to see whether the linear growth persists or saturates.

Load-bearing premise

The eigenvalues of the system matrix must be purely imaginary and the initial condition or parameters must satisfy a generic condition that lets the mean-motion argument apply.

What would settle it

A concrete counter-example would be any controllable linear system with purely imaginary eigenvalues together with a time-optimal control on an arbitrarily long interval whose number of switches stays bounded independently of interval length.

Figures

Figures reproduced from arXiv: 2502.13523 by Alexander Ovseevich, Michael Margaliot, Omri Dalin.

Figure 1
Figure 1. Figure 1: Geometric representation of zpφq “ zpφ1, φ2q. Since 1Bn pxq is a radial function (that is, it only depends only on |x|), ˆ1Bn pξq will depend on |ξ| (see, e.g. [5, Chapter IV]) and it is enough to consider the particular case where ξ “ “ 0 . . . 0 r ‰J , with r ą 0. Then, ˆ1Bn pξq “ 1 p2πq n{2 ż |x|ď1 e ´ixnr dx1 . . . dxn “ Vn´1 p2πq n{2 ż 1 ´1 p a 1 ´ x 2 n q n´1 e ´ixnr dxn, where Vn´1 is the volume of … view at source ↗
Figure 2
Figure 2. Figure 2: Plotting z “ a1e iφ1 ` a2e iφ2 , with φ1 “ 0, in the complex plane. We have |z| 2 “ ` a1 ` a2 cospφ2q ˘2 ` ` a2 sinpφ2q ˘2 “ a 2 1 ` 2a1a2 cospφ2q ` a 2 2 . Let qpr; a1, a2q :“ r 2 ´ pa 2 1 ` a 2 2 q 2a1a2 . (13) Then |z| 2 ď r 2 iff cospφ2q ď q, that is, iff arccospqq ď φ2 ď 2π ´ arccospqq. Eq. (12) now gives W2prq “ 1 p2πq 2 ż 2π 0 dφ1 ż 2π´arccospqq arccospqq dφ2 “ 1 2π ` 2π ´ 2 arccospqq ˘ “ 1 ´ arccos… view at source ↗
Figure 3
Figure 3. Figure 3: Plotting argpzpT qq T as a function of T. Then Ω “ ÿ 3 i“1 λiVi “ ? 2W2p1; 2.5, 3q ` 3W2p2.5; 1, 3q ` ? 3W2p3; 1, 2.5q “ ? 2 ż 8 0 J1pρqJ0p2.5ρqJ0p3ρqdρ ` 7.5 ż 8 0 J1p2.5ρqJ0pρqJ0p3ρqdρ ` 3 ? 3 ż 8 0 J1p3ρqJ0pρqJ0p2.5ρqdρ. Numerical integration using Matlab gives Ω “ 2.0614 (to 4-digit accuracy) [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
read the original abstract

We consider time-optimal controls of a controllable linear system with a scalar control on a long time interval. It is well-known that if all the eigenvalues of the matrix describing the linear system dynamics are real then any time-optimal control has a bounded number of switching points, where the bound does not depend on the length of the time interval. We consider the case where the governing matrix has purely imaginary eigenvalues, and show that then, in the generic case, the number of switching points is bounded from below by a linear function of the length of the time interval. The proof is based on relating the switching function in the optimal control problem to the mean motion problem that dates back to Lagrange and was solved by Hermann Weyl.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The manuscript considers time-optimal controls for controllable linear systems with scalar control over long time intervals. It recalls that real eigenvalues yield a bound on the number of switching points independent of interval length. For purely imaginary eigenvalues, it claims that in the generic case the number of switching points is bounded from below by a linear function of the interval length. The argument reduces the switching function to the mean-motion problem and applies Weyl's theorem.

Significance. If the central claim holds, the result is significant for optimal control theory: it establishes a sharp contrast between real and imaginary spectra regarding the growth of switching complexity. The reduction to Weyl's classical mean-motion result is a strength, supplying an external, non-fitted lower bound rather than a constructed quantity.

major comments (1)
  1. [main result section on generic case] The section presenting the main result and generic case: the reduction of the switching function to the mean-motion problem is sketched but the explicit verification that the generic condition produces a strictly positive rotation rate (hence the linear lower bound on zeros) is not fully detailed; this step is load-bearing for the claimed linear growth.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the positive evaluation of the significance of the result and for the constructive comment. We address the major comment below.

read point-by-point responses

  1. Referee: [main result section on generic case] The section presenting the main result and generic case: the reduction of the switching function to the mean-motion problem is sketched but the explicit verification that the generic condition produces a strictly positive rotation rate (hence the linear lower bound on zeros) is not fully detailed; this step is load-bearing for the claimed linear growth.

    Authors: We agree that the verification step could be expanded for clarity. The manuscript outlines the reduction of the switching function to a trigonometric polynomial and invokes the generic condition to ensure the rotation number is positive, but does not spell out the explicit computation of the mean motion under that condition. In the revised version we will insert a dedicated paragraph deriving that the generic non-resonance and non-vanishing amplitude assumptions imply a strictly positive rotation rate (via the explicit formula for the mean motion of a quasi-periodic function), from which Weyl's theorem directly yields the linear lower bound on the number of zeros. revision: yes

Circularity Check

0 steps flagged

No circularity; central bound obtained by applying external Weyl mean-motion theorem

full rationale

The derivation reduces the switching-function zero-counting problem to the classical mean-motion problem solved by Hermann Weyl (an external 1916 result independent of the present authors). The generic condition is invoked only to guarantee applicability of that external theorem and a positive rotation number, which then yields the linear lower bound. No parameter is fitted inside the paper and then renamed a prediction, no self-citation supplies a load-bearing uniqueness or ansatz, and the reduction is not self-definitional. The argument therefore remains self-contained against an external benchmark.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The proof rests on the external theorem of Weyl for the mean motion problem and on standard controllability assumptions for linear systems; no free parameters or invented entities are introduced in the abstract.

axioms (2)
  • standard math Weyl's solution to the mean motion problem
    Invoked to analyze the asymptotic behavior of the switching function over long intervals.
  • domain assumption The linear system is controllable
    Stated as a prerequisite for the time-optimal control problem.

pith-pipeline@v0.9.0 · 5646 in / 1200 out tokens · 31282 ms · 2026-05-23T02:50:15.306641+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

10 extracted references · 10 canonical work pages · 1 internal anchor

  1. [1]

    Athans and P

    M. Athans and P . L. Falb, Optimal Control: An Introduction to the Theory and Its Applications . McGraw-Hill Book Company, 1966

    work page 1966

  2. [2]

    Commentry on Minimax, Liapu nov and Bang-Bang,

    J. Nohel and H. J. Sussmann, “Commentry on Minimax, Liapu nov and Bang-Bang,” in Selected Papers of Norman Levinson , J. A. Nohel and D. H. Sattinger, Eds. Boston: Birkhauser, 1998, vol. 2, pp. 4 63–475

    work page 1998

  3. [3]

    Third-order nilpotency, ni ce reachability and asymptotic stability,

    Y . Sharon and M. Margaliot, “Third-order nilpotency, ni ce reachability and asymptotic stability,” J. Diff. Eqns. , vol. 233, no. 1, pp. 136–150, 2007

    work page 2007

  4. [4]

    E. D. Sontag, Mathematical Control Theory: Deterministic Finite Di- mensional Systems . New Y ork, NY: Springer, 1998

    work page 1998

  5. [5]

    E. M. Stein and G. Weiss, Introduction to F ourier Analysis on Euclidean Spaces. Princeton, NJ: Princeton University Press, 1990

    work page 1990

  6. [6]

    K. R. Stromberg, Introduction to Classical Real Analysis . Belmont, CA: Wadsworth Inc., 1981

    work page 1981

  7. [7]

    Reachability by means of nice controls,

    H. J. Sussmann, “Reachability by means of nice controls, ” in Proc. 26th IEEE Conf. on Decision and Control , vol. 26, 1987, pp. 1368–1373

    work page 1987

  8. [8]

    G. N. Watson, A Treatise on the Theory of Bessel Functions , 2nd ed. Camebridge, UK: Cambridge University Press, 1980

    work page 1980

  9. [9]

    Mean motion,

    H. Weyl, “Mean motion,” Amer . J Math., vol. 60, pp. 889–896, 1938

    work page 1938

  10. [10]

    An application of the mean motion problem to time-optimal control

    A. Wintner, “Upon a statistical method in the theory of d iophantine approximations,” Amer . J. Math., vol. 55, no. 1, pp. 309–331, 1933. This figure "W2FINAL.jpg" is available in "jpg" format from: http://arxiv.org/ps/2502.13523v1

    work page internal anchor Pith review Pith/arXiv arXiv 1933