Pith. sign in

REVIEW 4 major objections 6 minor 40 references

Rational Minimax Approximations for Matrix-Valued Functions: Existence, Optimality and Algorithms

T0 review · 4 major / 6 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read For continuous matrix-valued functions, rational minimax approximants with a shared denominator exist on point sets with no isolated points, and a positive-semidefinite Ruttan-type certificate guarantees global optimality; for sampled discr

desk verdict Strong continuum results for matrix-valued rational minimax, but the discrete certificate theory has a load-bearing existence gap. read the letter →

arxiv 2607.22576 v1 pith:HXDYUZ45 submitted 2026-06-06 math.OC cs.NAmath.NA

Lei-Hong Zhang , Chenkun Zhang This is my paper · ORCID
classification math.OCcs.NAmath.NA MSC 41A5241A5065D1549K3541A2090C46
keywords rationalminimaxapproximationmatrix-valuedfunctionscommondenominatorFrobeniusnormKolmogorovcriterionRuttanconditionstrongdualitym-d-Lawsoniteration
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 builds the missing theoretical floor for rational minimax approximation of matrix-valued functions whose entries share one common denominator, measured in the Frobenius norm. It proves, extending Walsh's scalar existence result, that a best approximant exists whenever the approximation set is dense in itself. It then derives necessary optimality conditions of Kolmogorov type and a sufficient Ruttan-type condition: if certain extreme error points and positive weights make a constructed Hermitian matrix positive semidefinite, the candidate is a global minimax approximant. For analytic functions on a continuum, the same certificate shows when boundary or discrete boundary-sample solutions also solve the original continuum problem. For finite node sets, the paper proves that strong duality in the dual formulation, including the practical condition that the dual denominator is nonzero at every node, is equivalent to the Ruttan certificate, and that the equations solved by the m-d-Lawson iteration are exactly the Kolmogorov dual criteria.

What carries the argument

The load-bearing object is the class R(s,t) of matrix-valued rational functions with a shared scalar denominator q(x) and entry numerators of prescribed degrees, with error measured by the Frobenius norm. The argument moves through three mechanisms: a Walsh-style diagonal extraction that preserves one common denominator to prove existence; the variational Kolmogorov conditions, which translate local optimality into inner-product inequalities at the extreme error points; and the Ruttan certificate matrix H = Σ ω_k H(z_k), a Hermitian block matrix built from the error at extreme points whose positive semidefiniteness forces global optimality. The discrete half of the paper identifies this cert

What would settle it

Find a finite sample X and continuous matrix-valued F where the dual maximizer w has an induced denominator q with q(x_j)=0 for some node but the error gap max_x ||F(x)-R(x)||^2 - d(w) is positive: that would directly falsify Theorem 5.2's equivalence without the denominator hypothesis. Even simpler: a finite X for which the infimum in (1.2) is not attained would falsify unlimited application of Theorem 2.1 to discrete problems.

Watch

Extended reading notes

Core claim

The central claim is that the matrix-valued rational minimax problem with common denominator admits a rigorous existence and optimality theory parallel to scalar rational approximation. The main theorems assert: existence of a minimax approximant on any set dense in itself; a primal Kolmogorov criterion and a dual Kolmogorov criterion characterizing local minimizers through extreme error points; a Ruttan-type sufficient condition for global optimality in which a weighted sum of Hermitian matrices over extreme points is positive semidefinite; a continuum-to-boundary transfer showing that a Ruttan certificate obtained on the boundary or a discrete subset certifies the continuum solution; and,

Load-bearing premise

The discrete-theory results assume that the finite-set minimax problem has a solution and that the dual optimizer's denominator is nonzero at every sampled node; the paper's existence theorem covers only point sets dense in themselves, so this assumption is unproven and the paper itself notes strong duality is not universally satisfied.

Editorial extensions

If this is right

  • If Theorem 2.1 holds, every continuous matrix-valued rational minimax problem on a self-dense set has an attained optimum, so numerical methods converge to a well-defined target rather than an infimum that may be missed.
  • If a Ruttan certificate is found for an analytic function on a Jordan domain, minimax approximants computed on the boundary or on a finite boundary sample containing the extreme points are automatically minimax on the whole domain.
  • If the computed dual solution in the discrete problem satisfies the nonzero-denominator condition and closes the duality gap, it is provably a global minimax approximant rather than merely a stationary point.
  • The orthogonality equations underlying m-d-Lawson coincide with the Kolmogorov dual criteria, so the iteration's weights and support points carry direct optimality meaning.
  • The duality gap can serve as a computable certificate of global optimality, replacing the impractical direct verification of the Ruttan matrix.

Reading between the lines

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

  • I infer that the common-denominator structure lets the matrix problem be decoupled once the denominator is fixed: for a fixed q, the numerators separate entrywise, so algorithms that alternate between denominator search and entrywise least-squares solves are natural consequences of this theory.
  • The paper leaves uniqueness, stability under boundary sampling, and discrete-to-continuum convergence open; a testable extension is to quantify how the Ruttan certificate degrades as boundary nodes are thinned, perhaps by monitoring the smallest eigenvalue of H.
  • The equivalence in Theorem 5.2 suggests that if the m-d-Lawson gap fails to close on a finite sample, the cause may be a vanishing dual denominator rather than a genuinely harder instance; reparameterizing or adding a regularization term that keeps q nonzero could restore strong duality.
  • Since Theorem 2.1 explicitly excludes finite sets, the discrete existence question is still open; one could test whether a limiting argument from dense sets containing the sample, or a barycentric parametrization, yields existence on finite X directly.
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

4 major / 6 minor

Summary. The paper develops a theoretical foundation for rational minimax approximation of continuous complex matrix-valued functions in the Frobenius norm, with all entries sharing a common denominator. The main contributions are: (1) an existence theorem for minimax approximants when the underlying point set is dense in itself (Theorem 2.1), extending Walsh's scalar theorem; (2) primal and dual Kolmogorov-type necessary conditions for local minimizers (Theorems 3.1 and 3.2); (3) a Ruttan-type sufficient condition for global optimality (Theorem 4.1) and its consequences for continuum, boundary, and discrete approximation (Theorem 4.2); and (4) an analysis of the discrete dual problem and the m-d-Lawson algorithm, claiming that strong duality is equivalent to Ruttan's condition and that the m-d-Lawson optimality equations coincide with Kolmogorov's dual criteria (Theorem 5.2 and Section 5.3). The paper is written in a classical approximation-theory style and explicitly acknowledges some limitations, such as the fact that Theorem 2.1 does not cover finite point sets and that the discrete quotient may not be well defined on the sampling set.

Significance. If the central claims are correct, this is a useful contribution. The existence theorem for matrix-valued common-denominator rational approximation on perfect sets is a genuine extension of Walsh's result, and the Ruttan-type certificate (Theorem 4.1) gives a practical, checkable sufficient condition for global optimality that could support computational methods. The paper also gives an honest discussion of the continuum-to-discrete passage, which is nontrivial even in the scalar case. The discrete duality analysis is potentially valuable for justifying algorithms such as m-d-Lawson. However, the manuscript has several load-bearing technical gaps: an unproved containment in Lemma 3.2, a non-Hermitian quadratic-form matrix in Section 4.1 due to missing conjugates, and a discrete existence gap that limits the scope of the advertised algorithmic certification theory. These issues do not appear to be merely cosmetic; they affect the proofs of Theorems 3.2, 4.1, and 5.2 as stated.

major comments (4)
  1. [§4.2 and §5 (Theorems 5.2 and the m-d-Lawson recovery)] The discrete part of the paper is load-bearing for the advertised algorithmic certification, but it is built on an unproven existence assumption. The paper states in §4.2 that Theorem 2.1 cannot be applied to the finite set X and defines R_X via 'arg inf' in (4.11). In §5, Theorem 5.2 and the recovery arguments require a dual optimizer with q(x_j) ≠ 0 on X, but no criterion is given for when such a pair exists. The issue is not academic: in the scalar case s=t=1, n=0, d=1, take X={0,1,2} and F(0)=0, F(1)=1, F(2)=0. Then m=3 satisfies (1.4), and the sequence R_c(x)=c/(1+x·(-1/(1+c))) has maximum error tending to 0, but no rational R∈R(0,1) attains error 0 because q(0)≠0 forces c=0, which makes R(1)=0. Thus the discrete minimax problem has no solution. Consequently, the equivalence in Theorem 5.2 and the m-d-Lawson recovery argument do not apply to this simple input. The abstract's caveat
  2. [§3.3, Lemma 3.2] The proof of Lemma 3.2 is incomplete at the crucial step. After defining the linear map T in (3.14), the proof says 'we obtain the desired result (3.13) by directly verifying that A ⊆ (T(Pν))⊥.' This verification is not shown, and for complex polynomial bases it is not evident. For a vector in A evaluated at x, the orthogonality to T(ξ) involves sums of the form (f_ij q - p_ij)(x) times inner products between the basis functions evaluated at x and the coefficients of ξ p_ij, together with analogous denominator terms. A direct expansion does not yield identically zero unless a specific basis and a specific conjugation convention are chosen and a nontrivial identity is proved. Since Lemma 3.2 is the sole justification for the subspace dimension bound (3.13) that feeds into Theorem 3.2's Carathéodory argument, the proof must be supplied. The same issue affects the consistency of Lemma 3.1,
  3. [§4.1, Eqs. (4.2)–(4.6)] The matrix H(x) defined in (4.3)–(4.6) is not Hermitian as written. In (4.4b), the (ℓ,u) entry of H_1^{(ij)}(x) is ψ_{ℓ-1}(x)ψ_{u-1}(x), and in (4.5) the entry of H_2(x) is (∥F(x)∥² - ∥ê∥)φ_{ℓ-1}(x)φ_{u-1}(x). For complex-valued basis functions, the quadratic form u(x;G,h)=∥F h - G∥² - |h|²∥ê∥ requires rank-one blocks built from ψ ψ^H and φ φ^H, not ψ ψ^T and φ φ^T. Without the conjugate on the second factor, H(x) is not Hermitian, so the condition H⪰0 in (4.7) is not well defined. The proof of Theorem 4.1 and its use in Theorem 4.2 and Theorem 5.2 all rely on this positive semidefinite matrix. This is a fixable but essential correction: the definitions must be made consistent with a fixed convention for coefficient vectors and polynomial evaluation.
  4. [§5.2, Theorem 5.2 and Abstract] The abstract claims that 'when the original minimax problem admits a solution, strong duality is equivalent to Ruttan's sufficient optimality condition.' The theorem as stated proves an equivalence between Ruttan's condition for a rational approximant and the equality d(ŵ)=∥ê∥_{∞,X} for a dual maximizer whose associated denominator is nonzero on X. This is not the same as a blanket equivalence between strong duality and Ruttan's condition. Moreover, the sufficiency direction of the proof invokes (5.18), a support-concentration statement from [40, Theorem 5.1], and uses it to assert that weights at non-extreme points vanish. Without a proof of (5.18) or a precise statement of the hypotheses under which it holds, Theorem 5.2 is conditional on an external result that is not included in the manuscript. The authors should either prove (5.18) or clearly state it as an assumption, and align the
minor comments (6)
  1. [Abstract] Typo: 'continuo us' should be 'continuous'.
  2. [Theorem 5.2 statement] The phrase 'Ruttan’s sufficient condition (4.1)' appears to refer to (4.7), since (4.1) defines the auxiliary function u(x;G,h). Please correct the equation reference.
  3. [§4.2, Eq. (4.11)] The notation 'R_X ∈ arg inf' is nonstandard when the infimum may not be attained. The paper already notes that Theorem 2.1 does not cover finite X, so the notation should be changed to something like 'inf' or the non-attainment should be explicitly acknowledged at this point.
  4. [§5.3] In the text following (5.19), the index range '0 ≤ ℓ0 ≤ n_{ij}' should be '0 ≤ ℓ0 ≤ n_{i0j0}' to match the fixed entry (i0,j0).
  5. [§3.3] In the proof of Lemma 3.1, the construction of G and h from the separating vector c involves a^H ψ and b^H φ; the relationship between this complex-conjugate pairing and the inner product ⟨Ê, ·⟩ in (3.3) needs to be spelled out. Even if Lemma 3.2 is repaired, this conjugation step should be made explicit for the complex case.
  6. [References] The paper relies heavily on the authors' own preprints [37] and [40], especially for the dual formulation, the m-d-Lawson algorithm, and the support-concentration result (5.18). Since these are not yet published, the manuscript should state which results are taken from those preprints and which are new here, and ideally make the needed external results precise.

Circularity Check

0 steps flagged · score 2.0 of 10

No construction-level circularity: the existence, optimality, and strong-duality proofs are genuine derivations from stated assumptions; the only caveat is reliance on the authors' own prior dual/Lawson framework and an explicitly conditional discrete-solvability hypothesis.

full rationale

I walked the main derivation chain. Theorem 2.1 is a Walsh-style compactness/continuity argument: after normalizing denominators, a diagonal extraction produces a common-denominator limit, and attainment is proved by contradiction against the minimizing sequence; attainability is not assumed. The Kolmogorov and Ruttan criteria (Theorems 3.1, 3.2, 4.1) are proved from first-order perturbations, Hahn-Banach/Carathéodory reduction, and a semidefinite quadratic-form inequality; the Ruttan sufficiency proof takes a hypothetical better approximant R̃ and derives a strict negativity contradicting H⪰0, so the global-optimality conclusion is not built into the hypothesis. Theorem 5.2's equivalence between Ruttan's condition and d(ŵ)=∥ê∥∞,X is a real two-way proof: sufficiency assembles H_ŵ=Σŵ_kH(x_k) from the generalized eigenproblem (5.9) and the support property; necessity constructs dual weights from the Ruttan certificate, uses weak duality and the pencil eigenvalue ordering to force d(ŵ)=η∞,X. Neither direction substitutes the conclusion for a premise. The abstract explicitly conditions the discrete strong-duality statement on 'when the original minimax problem admits a solution', and Section 5 notes Theorem 2.1 does not cover finite X; that is a stated hypothesis/gap, not a circular input. The paper cites the same authors' [37,40] for the scalar Ruttan-duality bridge, the dual formulation, and m-d-Lawson convergence; those are prior technical infrastructure rather than assumed instances of the new theorems. Thus no construction-level circularity is present; at most there is minor self-citation burden, which I score as 2 rather than 0.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The paper's central claims rest on standard approximation-theory tools (Walsh compactness, Hahn-Banach, Carathéodory, maximum norm principle) and on conditional assumptions for the discrete part. No free parameters or fitted constants are used, and no new physical or algorithmic entities are postulated.

assumptions (5)
  • standard math Walsh's scalar rational convergence theorem and compactness argument
    Used in Theorem 2.1; the proof invokes Walsh [31, Theorems I, II] for subsequential continuous convergence of rational functions.
  • standard math Maximum Frobenius norm principle for analytic matrix-valued functions
    Used in §1 and Theorem 4.2 to move from domain D to boundary L; for F∈CA(D), R pole-free in D, the maximum error occurs on ∂D.
  • standard math Hahn-Banach separation and Carathéodory's lemma
    Used in Lemma 3.1/3.3 and Theorem 3.2 to derive the dual Kolmogorov criterion.
  • domain assumption Discrete minimax problem admits a solution and denominator is nonzero at all sample nodes
    Explicitly assumed in §5/Theorem 5.2; Theorem 2.1 does not cover finite X, so this is a load-bearing assumption for the algorithm/certificate results.
  • domain assumption Irreducible representation gcd(P,q)=1 with defect ν
    Needed to define unique representation and defect ν in §3.1; any rational approximant can be put in this form.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Rational Minimax Approximations for Matrix-Valued Functions: Existence, Optimality and Algorithms." pith.science (2026). https://pith.science/paper/HXDYUZ45

@misc{pith2026260722576,
  author       = {Pith},
  title        = {Pith review of: Rational Minimax Approximations for Matrix-Valued Functions: Existence, Optimality and Algorithms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HXDYUZ45}},
  note         = {Machine review of arXiv:2607.22576}
}
read the original abstract

In this paper, we study rational minimax approximation for continuous complex matrix-valued functions in the Frobenius norm, where all approximant entries share a common denominator. This generalizes classical scalar rational approximation, with applications in system modeling, microwave design, and nonlinear eigenvalue problems. We first prove the existence of such matrix-valued approximants on point sets dense in themselves, extending Walsh's foundational scalar result. Next, we establish characterizations of the local and global minimax approximants by deriving primal/dual matrix-valued Kolmogorov criteria and a Ruttan-type sufficient condition for global optimality. For analytic functions on a continuum, we link continuum minimax approximation to approximation on its boundary, and finite boundary samples via the maximum norm principle. We show that Ruttan's sufficient optimality condition provides a certificate under which a minimax approximant obtained from the boundary or from a discrete set of boundary nodes also solves the original continuum problem. Finally, for discrete approximation, we connect these conditions to a dual problem and the related dual-based numerical method m-d-Lawson: when the original minimax problem admits a solution, strong duality is equivalent to Ruttan's sufficient optimality condition, and, the optimality equations underlying the m-d-Lawson iteration coincide with Kolmogorov's dual criteria. These results provide a theoretical basis for certifying and computing matrix-valued rational minimax approximants.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 2 canonical work pages

  1. [1]

    URL https://history-of-approximation-theory.com

  2. [2]

    Berljafa and S

    M. Berljafa and S. G¨ uttel, The RKFIT algorithm for nonlinear rat ional approximation, SIAM J. Sci. Comput. , 39 (2017), A2049–A2071, URL https://doi.org/10.1137/15M1025426

    doi:10.1137/15m1025426 2017

  3. [3]

    E. W. Cheney, Introduction to Approximation Theory , 2nd edition, Chelsea Publishing Company, New York, 1982

    1982

  4. [4]

    A. A. Condori, Maximum principles for matrix-valued analytic funct ions, Amer. Math. Monthly, 127 (2020), 331–343. 25

    2020

  5. [5]

    Deschrijver, M

    D. Deschrijver, M. Mrozowski, T. Dhaene and D. de Zutter, Mac romodeling of multiport systems using a fast implementation of the vector fitting method, IEEE Microw. Wireless Compon. Lett., 18 (2008), 383–385

    2008

  6. [6]

    T. A. Driscoll, Y. Nakatsukasa and L. N. Trefethen, AAA rationa l approximation on a continuum, SIAM J. Sci. Comput. , 46 (2024), A929–A952

    2024

  7. [7]

    I. V. Gosea and S. G¨ uttel, Algorithms for the rational approxim ation of matrix-valued functions, SIAM J. Sci. Comput. , 43 (2021), A3033–A3054, URL https://doi.org/10.1137/20M1324727

    doi:10.1137/20m1324727 2021

  8. [8]

    Gustavsen, Improving the pole relocating properties of vect or fitting, IEEE Trans

    B. Gustavsen, Improving the pole relocating properties of vect or fitting, IEEE Trans. Power Deliv., 21 (2006), 1587–1592

    2006

Show all 40 references
  1. [9]

    Gustavsen, Matrix Fitting Toolbox, 2009, Https://www.sintef .no/en/software/vector- fitting/downloads/matrix-fitting-toolbox/

    B. Gustavsen, Matrix Fitting Toolbox, 2009, Https://www.sintef .no/en/software/vector- fitting/downloads/matrix-fitting-toolbox/

    2009

  2. [10]

    Gustavsen and A

    B. Gustavsen and A. Semlyen, Rational approximation of frequ ency domain responses by vector fitting, IEEE Trans. Power Deliv. , 14 (1999), 1052–1061

    1999

  3. [11]

    M. H. Gutknecht, On complex rational approximation. Part I: T he characterization problem, in Computational Aspects of Complex Analysis (H. Werneret at. ,eds.). Dordrecht : Reidel , 1983, 79–101

    1983

  4. [12]

    G¨ uttel and F

    S. G¨ uttel and F. Tisseur, The nonlinear eigenvalue problem, Acta Numer. , 26 (2017), 1–94

    2017

  5. [13]

    G¨ uttel, G

    S. G¨ uttel, G. M. Negri Porzio and F. Tisseur, Robust rational approximations of non- linear eigenvalue problems, SIAM J. Sci. Comput. , 44 (2022), A2439–A2463, URL https://doi.org/10.1137/20M1380533

    2022 doi

  6. [14]

    A. N. Kolmogorov, A remark on the polynomials of Chebyshev dev iating least from a given function, Uspekhi Mat. Nauk. , 3 (1948), 216–221 (in Russian)

    1948

  7. [15]

    Lietaert, K

    P. Lietaert, K. Meerbergen, J. P´ erez and B. Vandereycken , Automatic rational approximation and linearization of nonlinear eigenvalue problems, IMA J. Numer. Anal. , 42 (2022), 1087– 1115

    2022

  8. [16]

    Meinardus, Approximation of Functions: Theory and Numerical Methods

    G. Meinardus, Approximation of Functions: Theory and Numerical Methods. Translated by L. L. Schumaker. , Berlin Heidelberg, Springer, 1967

    1967

  9. [17]

    Nakatsukasa, O

    Y. Nakatsukasa, O. S` ete and L. N. Trefethen, The first five years of the AAA algorithm, Intl. Cong. Basic Sci. , URL https://arxiv.org/pdf/2312.03565, To appear

    arXiv

  10. [18]

    Nakatsukasa and L

    Y. Nakatsukasa and L. N. Trefethen, Applications of AAA ratio nal approximation, Acta Numer., to appear, URL https://arxiv.org/abs/2510.16237

    arXiv

  11. [19]

    J. R. Rice, The approximation of functions , Addison-Wesley publishing Company, 1969

    1969

  12. [20]

    T. J. Rivlin and H. S. Shapiro, A unified approach to certain proble ms of approximation and minimization, J. Soc. Indust. Appl. Math. , 9 (1961), 670–699

    1961

  13. [21]

    Ruttan, A characterization of best complex rational appro ximants in a fundamental case, Constr

    A. Ruttan, A characterization of best complex rational appro ximants in a fundamental case, Constr. Approx. , 1 (1985), 287–296. 26

    1985

  14. [22]

    Y. Saad, M. El-Guide and A. Mi¸ edlar, A rational approximation method for the nonlinear eigenvalue problem, Technical report, 2020, URL https://arxiv.org/abs/1901.01188v2

    2020 arXiv

  15. [23]

    H. S. Shapiro, Topics in Approximation Theory , Lecture Notes in Mathematics, Springer- Verlag, Berlin, Heidelberg, 1971

    1971

  16. [24]

    Singer, Best approximation in normed linear spaces by elements of li near subspaces, Springer- Verlag, Berlin-Heidelberg-New York, 1970

    I. Singer, Best approximation in normed linear spaces by elements of li near subspaces, Springer- Verlag, Berlin-Heidelberg-New York, 1970

    1970

  17. [25]

    Singer, Duality for Nonconvex Approximation and Optimization , Springer-Verlag, New York, NY, 2006, URL https://doi.org/10.1007/0-387-28395-1

    I. Singer, Duality for Nonconvex Approximation and Optimization , Springer-Verlag, New York, NY, 2006, URL https://doi.org/10.1007/0-387-28395-1

    2006 doi

  18. [26]

    Thiran and M.-P

    J.-P. Thiran and M.-P. Istace, Optimality and uniqueness conditio ns in complex rational Chebyshev approximation with examples, Constr. Approx. , 9 (1993), 83–103

    1993

  19. [27]

    Traina, Cauchy Method Modeling: Applied to Computer-Aided Tuning o f Microwave Filters and Duplexers , VDM Verlag Dr

    D. Traina, Cauchy Method Modeling: Applied to Computer-Aided Tuning o f Microwave Filters and Duplexers , VDM Verlag Dr. M¨ uller, 2010

    2010

  20. [28]

    Traina, G

    D. Traina, G. Macchiarella and T. K. Sarkar, Robust formulation s of the Cauchy method suitable for microwave duplexers modeling, IEEE Trans. Microw. Theory Tech. , 55 (2007), 974–982

    2007

  21. [29]

    L. N. Trefethen, Approximation Theory and Approximation Practice, Extende d Edition, SIAM, 2019

    2019

  22. [30]

    L. N. Trefethen and M. H. Gutknecht, Real vs. complex ration al Chebyshev approximation on an interval, Trans. Amer. Math. Soc. , 280 (1983), 555–561

    1983

  23. [31]

    J. L. Walsh, The existence of rational functions of best appro ximation, Trans. Amer. Math. Soc., 33 (1931), 668–689

    1931

  24. [32]

    J. L. Walsh, Interpolation and Approximation by Rational Functions in t he Complex Domain , 5th edition, Amer. Math. Soc. Colloq. Publ., vol 20, R.I. USA, 1969

    1969

  25. [33]

    Williams, Characterization and computation of rational Cheby shev approximations in the complex plane, SIAM J

    J. Williams, Characterization and computation of rational Cheby shev approximations in the complex plane, SIAM J. Numer. Anal. , 16 (1979), 819–827

    1979

  26. [34]

    Zhang, J

    C. Zhang, J. Gu and L.-H. Zhang, NEP MiniMax: An approach for NEPs based on matrix- valued minimax approximations, URL https://arxiv.org/abs/2603.13794

  27. [35]

    Zhang and S

    L.-H. Zhang and S. Han, A convergence analysis of Lawson’s iter ation for computing polyno- mial and rational minimax approximations, SIAM J. Numer. Anal. , 63 (2025), 2249–2271

    2025

  28. [36]

    Zhang, L

    L.-H. Zhang, L. Yang, W. H. Yang and Y.-N. Zhang, A convex dua l problem for the rational minimax approximation and Lawson’s iteration, Math. Comp. , 94 (2025), 2457–2494, DOI: https://doi.org/10.1090/mcom/4021

    2025 doi

  29. [37]

    Zhang, Optimality conditions for rational minimax approxima tions: Bridging Ruttan’s criteria to dual-based methods, 2026, URL https://arxiv.org/abs/2602.07862v1

    L.-H. Zhang, Optimality conditions for rational minimax approxima tions: Bridging Ruttan’s criteria to dual-based methods, 2026, URL https://arxiv.org/abs/2602.07862v1

    2026

  30. [38]

    Zhang and Y

    L.-H. Zhang and Y. Zhang, Interpolation constrained rational minimax approx- imation with barycentric representation, J. Comput. Math. , to appear, URL https://arxiv.org/pdf/2502.10665v2. 27

    arXiv

  31. [39]

    Zhang, Y.-N

    L.-H. Zhang, Y.-N. Zhang, L. Yang and R.-W. Xiao, A AAA-type alg orithm for the microwave duplexer filtering, Numer. Algorithms , 101 (2025), 1607–1631, Https://doi.org/10.1007/s11075-025-02058-0

    2025 doi

  32. [40]

    Zhang, Y.-N

    L.-H. Zhang, Y.-N. Zhang, C. Zhang and S. Han, Rational minimax approximation of matrix- valued functions, 2025, URL https://arxiv.org/pdf/2508.06378v2. 28

    2025 arXiv

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.