Pith. sign in

REVIEW 4 minor 40 references

Synthesis of passive multidimensional scattering matrices: a survey

T0 review · 0 major / 4 minor · reviewed 2026-07-12 · grok-4.5

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

Pith's one-line read Every 2-D lossless passive scattering matrix can be built from two kinds of reactive elements; higher dimensions fail because positive polynomials are not always sums of squares.

desk verdict Solid specialist survey that cleanly maps the 2-D frontier and the SOS obstruction for n>2; real-coefficient spectral factors remain open but are already flagged by the author. read the letter →

arxiv 2607.02557 v1 pith:AADBEWYV submitted 2026-06-27 eess.SP

Sankar Basu (National Science Foundation , USA) This is my paper · ORCID
classification eess.SP MSC 94C0593B5014P1015A23
keywords multidimensionalnetworksynthesispassivescatteringmatricessumofsquaresHilbert17thproblemspectralfactorizationbounded-reallemmalosslessembedding2-Dsystems
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 survey shows how far classical one-dimensional network synthesis can be pushed into several variables. In two dimensions every rational lossless bounded scattering matrix is realizable by a finite network of two types of inductors or capacitors plus memoryless multiports, and the construction is minimal in the McMillan-degree sense. The same path works for dissipative (lossy) 2-D matrices only after they are embedded into a larger lossless matrix; that embedding rests on a partial spectral-factorization result that itself rests on Hilbert’s theorem that positive bivariate polynomials are sums of four rational squares. In three or more variables even scalar lossless all-pass functions cease to be synthesizable except for certain low-degree cases, again because positive polynomials lack adequate sum-of-squares representations. The paper therefore maps the precise boundary between what passive multidimensional synthesis can and cannot achieve, and identifies Hilbert’s 17th problem as the fundamental obstruction.

What carries the argument

The 2-D matrix spectral factorization theorem (Theorem 2.2): every para-Hermitian polynomial matrix nonnegative on the imaginary axes admits a holomorphic factor of size 2r imes m whose least common denominator is univariate; this factorization, obtained via Cassel reduction of Hilbert’s sum-of-squares representation, supplies both the unitary embedding of dissipative matrices and the building blocks of the lossless synthesis algorithm.

What would settle it

Exhibit a concrete real-coefficient 2-D para-Hermitian positive polynomial matrix for which every holomorphic spectral factor of the form given by Theorem 2.2 necessarily has non-real coefficients, or exhibit a real 3-D degree-one all-pass that cannot be realized by any finite passive network of three reactive types.

Watch

Extended reading notes

Core claim

The central claim is that the classical synthesis of passive scattering matrices extends completely to two variables for lossless systems and only partially for dissipative systems, but collapses for three or more variables; the collapse is forced by the failure of positive polynomials to be sums of squares of polynomials (or of polynomials with univariate denominators) once the number of variables exceeds two.

Load-bearing premise

The constructive spectral factors and embeddings are allowed to have complex coefficients even when the original transfer matrix has real coefficients, so the networks that are produced may require complex-valued elements.

Editorial extensions

If this is right

  • Any rational 2-D lossless scattering matrix can be realized with exactly ν1 of one reactive type and ν2 of the other, where the νi are the McMillan degrees read from the determinant.
  • Dissipative 2-D bounded matrices become synthesizable once they are unitarily embedded; the number of resistive ports needed is twice the normal rank of I−S*S.
  • No general synthesis procedure exists for n>2, even for scalar all-pass functions of total degree greater than one in each variable.
  • Cascade or cascade-lattice factorizations that work for 1-D two-ports fail generically in two or more variables because the associated linear systems for the Belevitch polynomials become over-determined.
  • A weak 2-D bounded-real lemma holds for the particular Roesser realizations obtained from the synthesis procedure, but state-space isomorphism fails so the lemma does not extend to arbitrary minimal realizations.

Reading between the lines

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

  • If a constructive algorithm that forces real spectral factors could be found, real-element 2-D passive synthesis would become complete; the paper leaves this sharpening of the sum-of-squares step open.
  • The same SOS obstruction that blocks n-D synthesis also limits the existence of outer spectral factors needed for multidimensional Wiener filtering and hyperstability theory.
  • The low-degree 3-D all-passes that do synthesize may seed a recursive construction of larger synthesizable families once a suitable inductive step is identified.
  • Numerical SDP certificates of positivity could supply approximate spectral factors even when exact rational factors do not exist, opening a route to approximate n-D passive design.
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

0 major / 4 minor

Summary. The paper surveys the status of synthesizing rational n-D passive (bounded-real or positive-real) transfer functions as scattering or immittance matrices of networks built from n types of reactive elements plus memoryless multiports. It shows that every 2-D lossless bounded matrix is synthesizable by a three-step construction (extraction of a 1-D coupling network, similarity transformation via Lyapunov factorization of a univariate kernel K, and degree reduction) that is minimal in the McMillan degrees ν1, ν2 (Section 3). Dissipative 2-D matrices can be unitarily embedded into lossless ones via a 2-D matrix spectral factorization that rests on Cassel reduction of bivariate positive polynomials (Theorems 2.2 and 5.1), after which the lossless synthesis applies; the embedding is only partial and fails for n>2. For n>2 even scalar lossless all-passes are generically non-synthesizable, the sole positive exception being certain degree-one-in-each-variable 3-D all-passes that exploit Hilbert’s theorem on ternary quartics (Section 4). All limitations are traced to the absence of adequate sum-of-squares representations (Hilbert’s 17th problem). Alternative cascade and elementary-symmetric constructions are also reviewed.

Significance. The survey supplies a coherent, self-contained account of a classical open problem that has remained fragmented across circuit theory, multidimensional systems, and real algebraic geometry for half a century. By making the precise link between network synthesizability and Hilbert’s 17th problem explicit, and by collecting the constructive 2-D algorithms (Smith form over C(p1)[p2], Cassel reduction, Lyapunov factorization of K, Belevitch-type embedding for low-degree 3-D all-passes) together with the known counter-examples, the paper becomes a definitive reference for both engineers and mathematicians working on multidimensional passive systems. The constructive sketches are detailed enough for a survey and correctly flag the remaining real-coefficient spectral-factor gap.

minor comments (4)
  1. Several cross-references are incomplete or placeholder-like (e.g., “[9, Theorem xxx]” in Lemma 5.1 and the surrounding text). These should be replaced by the actual theorem numbers from the cited papers before final publication.
  2. The continuous-to-discrete translation of the spectral factor (remark after Theorem 2.3) leaves open whether a univariate denominator can always be retained; a short clarifying sentence would help readers who work exclusively in the discrete domain.
  3. Notation for the discrete paraconjugate toggles between tilde and asterisk in a few places (e.g., around (3.15)–(3.17)); a uniform convention would improve readability.
  4. Table 1 (feasibility of SOS representations) is useful but could briefly note that the n=2, d=4 entry is Hilbert’s classical ternary-quartic result, which is later invoked for the 3-D degree-one all-pass construction.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: survey sketches constructive algebraic derivations from classical external SOS/Smith/1-D results; self-citations are to prior parameter-free algorithms, not load-bearing reductions.

full rationale

The paper is a self-contained specialist survey whose central status claims (full 2-D lossless synthesizability via the three-step construction of §3 that produces a minimal discrete lossless Sf; partial 2-D dissipative unitary embedding via Theorems 2.2/5.1 that rest on Cassel-reduced bivariate SOS + Smith form over the Euclidean domain C(p1)[p2]; generic failure for n>2 via Hilbert’s 17th and the concrete counter-example of §4.1.2; and the low-degree 3-D all-pass exception of §4.1.1 that uses Hilbert’s ternary-quartic theorem) are derived by explicit algebraic constructions and classical external theorems (Hilbert, Artin–Schreier, Pfister, Smith normal form, Oono–Yasuura, Youla 1-D factorization, Fettweis–Basu scattering-Hurwitz theory). No step reduces a claimed prediction or first-principles result to its own input by definition, by a fitted parameter, or by an unverified self-citation chain. Self-citations (e.g., to the author’s earlier constructive spectral-factorization algorithm [8] or to joint work with Kummert) supply historical context or alternative presentations of the same algebraic steps already sketched in the present text; they are not the sole justification of any load-bearing claim. There is no data-fitting, no uniqueness theorem imported solely from the author’s prior papers, no ansatz smuggled via citation, and no renaming of a known empirical pattern. The acknowledged gap that spectral factors need not be real-coefficient (remarks after Theorem 2.2 and §8) is already flagged by the authors and does not create a circular loop. Score 1 reflects only the presence of ordinary self-citation that is not load-bearing.

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

Pure mathematical survey. No free parameters fitted to data. Load-bearing background consists of standard algebraic and network-theoretic facts plus a few domain conventions of multidimensional circuit theory. No new physical entities are postulated.

assumptions (6)
  • standard math Hilbert’s theorem on ternary quartics and Artin–Schreier solution of Hilbert’s 17th problem: a positive polynomial in n variables is a sum of squares of rational functions (with Pfister bound π(n) ≤ 2^n); bivariate positive polynomials admit a Cassel reduction to a sum of four squares with univariat
    Invoked throughout §2 as the foundation of 2-D spectral factorization and the obstruction for n>2.
  • standard math C(p1)[p2] (and R(p1)[p2]) is a Euclidean domain / PID, so every matrix admits a Smith normal form with monic invariant factors.
    Fact 2.4; used to reduce the matrix spectral-factorization problem.
  • standard math Classical 1-D spectral factorization of para-Hermitian nonnegative rational matrices and the Oono–Yasuura factorization of p2-unimodular para-Hermitian nonnegative matrices (Fact 2.5).
    Used after Kronecker-product reduction to finish the 2-D matrix factorization.
  • domain assumption Scattering-Hurwitz (scattering-Schur) polynomials are closed under certain specializations and products; a rational matrix is lossless bounded iff it is holomorphic in the right half polyplane (open polydisc) and para-unitary on the boundary.
    Standard multidimensional network theory (Fettweis–Basu); used to guarantee holomorphy of constructed factors and of the coupling network Sf.
  • domain assumption Every 1-D discrete lossless bounded matrix admits a minimal passive realization with a number of delays equal to the McMillan degree of its determinant (standard Belevitch / Youla synthesis).
    Used in §3.3–3.4 to realize the univariate coupling network Sf and to prove minimality of the 2-D count ν1, ν2.
  • domain assumption Roesser state-space model is an adequate realization framework for the weak 2-D bounded-real lemma; block-diagonal similarity is the natural notion of equivalence (even though full state-space isomorphism fails in 2-D).
    §6; the weak BR lemma is stated only for the particular passive realization obtained from synthesis, not for arbitrary minimal realizations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Synthesis of passive multidimensional scattering matrices: a survey." pith.science (2026). https://pith.science/paper/AADBEWYV

@misc{pith2026260702557,
  author       = {Pith},
  title        = {Pith review of: Synthesis of passive multidimensional scattering matrices: a survey},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AADBEWYV}},
  note         = {Machine review of arXiv:2607.02557}
}
abstract

The goal of this exposition is twofold. First, it surveys the current status of the synthesis of rational multivariable ($n$-dimensional) passive linear shift-invariant bounded-real (or positive-real) transfer functions as scattering (or immittance) matrices of networks composed of a finite number of $n$ types of inductive and capacitive elements together with memoryless reciprocal and nonreciprocal components, such as transformers and gyrators. The theory seeks to extend the classical $1$-D network synthesis framework to the multidimensional ($n$-D) setting. Second, the exposition examines the fundamental connection between $n$-D network synthesis and Hilbert's 17th problem concerning the representation of positive polynomials as sums of squares. This connection is shown to be the principal mathematical obstacle to extending many classical synthesis results to higher dimensions. Both lossless and dissipative transfer functions are considered. While every $2$-D lossless transfer function can be synthesized, the classical procedure of embedding a dissipative transfer function into a lossless one by solving an associated matrix dilation problem can be executed via spectral factorization only partially in $2$-D and fails in general for $n>2$. Moreover, for $n>2$, even lossless transfer functions are not, in general, synthesizable, except for certain low-degree stable all-pass functions. These limitations ultimately stem from Hilbert's 17th problem referred to above. The exposition also reviews failure of several other classical $1$-D synthesis techniques, including matrix-factorization methods, thereby providing a comprehensive account of the current state of $n$-D network synthesis. Depending on the context, both continuous and discrete time formulations are employed; for the class of linear shift-invariant systems considered here, this change of setting entails no loss of generality.

Figures

Figures reproduced from arXiv: 2607.02557 by the authors.

Figure 1
Figure 1. Steps in synthesis of a bounded lossless scattering matrix. [PITH_FULL_IMAGE:figures/full_fig_p028_1.png] view at source ↗
Figure 2
Figure 2. Proof of bounded real lemma. Input-output variables [PITH_FULL_IMAGE:figures/full_fig_p053_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 1 canonical work pages

  1. [1]

    Ann., vol

    David Hilbert, ¨Uber die Darstellung definiter Formen als Summe von Formengnadraten, Math. Ann., vol. 32 (1888), pp. 342-350

  2. [2]

    David Hilbert, ¨Uber Tern¨ are Definite Formen, Acta Mathematica, vol. 32, pp. 169-197, 1893. 60

  3. [3]

    Landau, ¨Uber die Darstellung definiter Funktionen durch Quadrate, Mathematische Annalen, vol

    E. Landau, ¨Uber die Darstellung definiter Funktionen durch Quadrate, Mathematische Annalen, vol. 62, p. 272, 1906

    1906

  4. [4]

    T. S. Motzkin, The arithmatic-geometric inequality, in Inequalities, Oved Shisha (ed.), Academic Press, New York, 1967

    1967

  5. [5]

    Youla, On the factorization of rational matrices, IEEE Trans

    D.C. Youla, On the factorization of rational matrices, IEEE Trans. on IT, Vol. 18, No. 7, pp 172-189, 1961

    1961

  6. [6]

    D. C. Youla, The analysis and synthesis of lumped passive n- dimensional networks; part I: Analysis, Tech Rep. Polytech. Inst. New York, Poly- MRI-1437-84

  7. [7]

    Koga, Synthesis of finite passiven-ports with prescribed two variable reactance matrices, IEEE Trans

    T. Koga, Synthesis of finite passiven-ports with prescribed two variable reactance matrices, IEEE Trans. on Circuit Theory, pp. 31-52, March 1966

    1966

  8. [8]

    Basu, A constructive algorithm for 2-D spectral factorization with rational spectral factors, IEEE Transactions on Circuits and Systems - Part I, vol

    S. Basu, A constructive algorithm for 2-D spectral factorization with rational spectral factors, IEEE Transactions on Circuits and Systems - Part I, vol. 47, no. 9, pp. 1309-1318, Sept. 2000

    2000

Show all 40 references
  1. [9]

    Fettweis and S

    A. Fettweis and S. Basu, New results on stable multidimensional poly- nomials, part I: continuous case, IEEE Transactions on Circuits and Systems, vol. 34, no. 10, pp. 1221-1232, October 1987

    1987

  2. [10]

    Fettweis, Elementary symmetric functions and positive functions of several variables, Proc

    A. Fettweis, Elementary symmetric functions and positive functions of several variables, Proc. of Int. Symp. on Mathematical Theory of Net- works and Systems, pp. 91-98, vol III, 1989

    1989

  3. [11]

    Basu and A

    S. Basu and A. Fettweis, New results on stable multidimensional polyno- mials, part II: discrete case, IEEE Transactions on Circuits and Systems, vol. 34, no. 11, pp. 1264-1275, November 1987

    1987

  4. [12]

    Basu, New results on stable multidimensional polynomials, Part III: State Space Interpretations IEEE Transactions on Circuits and Systems, vol

    S. Basu, New results on stable multidimensional polynomials, Part III: State Space Interpretations IEEE Transactions on Circuits and Systems, vol. 38, no. 7, pp. 755-768, July 1991

    1991

  5. [13]

    Basu, A framework for 2-D hyperstability theory based provably con- vergent adaptive 2-D IIR filetring, SIAM J

    S. Basu, A framework for 2-D hyperstability theory based provably con- vergent adaptive 2-D IIR filetring, SIAM J. on Control & Opt., vol. 29, no. 6, pp. 1476-1508, November 1991. 61

    1991

  6. [14]

    Basu and A

    S. Basu and A. Tan, On a generalized factorization problem for the structurally passive synthesis of multidimensional digital filters, IEEE Transactions on Circuits and Systems, vol.36, no.12, pp. 1529-1544, De- cember 1989

    1989

  7. [15]

    Basu, Multi-dimensional filter banks and wavelets: a system theo- retic perspective, J

    S. Basu, Multi-dimensional filter banks and wavelets: a system theo- retic perspective, J. of the Franklin Institute, vol. 3335B, pp. 1367-1409, November 1998

    1998

  8. [16]

    253, 2000

    Bruce Reznick, Some concrete aspects of Hilbert’s 17th problem, Con- temporary Mathematics, vol. 253, 2000. DOI: 10.1090/conm/253/03936

    2000 doi

  9. [17]

    T. Y. Lam,Introduction to Quadratic forms over Fields, American Mathematical Society, Prividence, Rhode Island, 2005

    2005

  10. [18]

    Grigoriy Blekherman, There are significantly more nonegative polyno- mials than the sums of squares, Israel Journal of Mathematics, vol. 153, pp. 355-380, 2006

    2006

  11. [19]

    Lasserre, A Sum of Squares Approximation of Nonnegative Poly- nomials, SIAM review, pp.651-669, vol

    Jean B. Lasserre, A Sum of Squares Approximation of Nonnegative Poly- nomials, SIAM review, pp.651-669, vol. 49, no. 4, 2007

    2007

  12. [20]

    N. K. Bose,Applied multidimensional systems theory, Van Nostrand Reinhold, 1982

    1982

  13. [21]

    146, 2000

    Murray Marshall,Positive Polyomials and Sums of Squares, AMS Math- ematical Surveys and Monograpghs, vol. 146, 2000

    2000

  14. [22]

    Peter Cameron,An Introduction to Algebra, Oxford University Press, 2008

    2008

  15. [23]

    Belevitch,Classical Network Theory, Holden-Day, San Fransisco, 1968

    V. Belevitch,Classical Network Theory, Holden-Day, San Fransisco, 1968

    1968

  16. [24]

    Kailath,Linear Systems Theory, Prentice Hall, 1982

    T. Kailath,Linear Systems Theory, Prentice Hall, 1982

    1982

  17. [25]

    Oono, and K

    Y. Oono, and K. Yasuura, Synthesis of finite passive n-terminal networks with prescribed scattering matrices, Memorandum of Kyushu University (Engineering), Japan, pp. 125-177, vol. 14., No. 2, 1954. 62

    1954

  18. [26]

    Kummert, The synthesis of two-dimensional passive n-ports contain- ing lumped elements, J

    A. Kummert, The synthesis of two-dimensional passive n-ports contain- ing lumped elements, J. of Multidimensional Systems and Signal Pro- cessing, vol. 1, no.2, pp. 351-362, 1990

    1990

  19. [27]

    Kummert, On the synthesis of multidimensional reactance multi- ports, IEEE Transactions on Circuits and Systems, pp

    A. Kummert, On the synthesis of multidimensional reactance multi- ports, IEEE Transactions on Circuits and Systems, pp. 637-642, vol. 38, no. 6, June 1991

    1991

  20. [28]

    Kummert, The synthesis of 3-D lossless first-order one ports with lumped elements, IEEE Trans

    A. Kummert, The synthesis of 3-D lossless first-order one ports with lumped elements, IEEE Trans. on Circuits and Systems, pp. 1445-1448, vol. 36, no. 11, November 1989

    1989

  21. [29]

    Kummert, Spectral factorization of two-variable para-Herminitian polynomial matrices, Multidimensional Systems and Signal Processing, vol

    A. Kummert, Spectral factorization of two-variable para-Herminitian polynomial matrices, Multidimensional Systems and Signal Processing, vol. 1, no.2, pp. 327-339, 1990

    1990

  22. [30]

    Fettweis and A

    A. Fettweis and A. Kummert, An efficient algorithms for the spectral factorization of rational nonnegative parahermitian matrices, AE ¨U, vol 46, no 3, pp. 150-156, 1992

    1992

  23. [31]

    Vidyasagar,Control System Synthesis - a Factorization Approach, MIT Press, 1986

    M. Vidyasagar,Control System Synthesis - a Factorization Approach, MIT Press, 1986

    1986

  24. [32]

    Dewilde and E

    P. Dewilde and E. Deprettere, Orthogonal cascade realizations of real multiport digital filters, Int. J. Circuit Th. & Appl., vol. 8, pp. 245- 277, 1980

    1980

  25. [33]

    Dewilde, V

    P. Dewilde, V. Belevitch and R. W. Newcomb, The problem of degree reduction of a scattering matrix by factorization, J. of the Franklin In- stitute, vol. 291, no. 5, May 1971

    1971

  26. [34]

    Vandewalle and Patrick Dewilde, On the Minimal Spectral Fac- torization of Nonsingular Positive Rational Matrices, IEEE Transactions on Information Theory, vol

    Joos P. Vandewalle and Patrick Dewilde, On the Minimal Spectral Fac- torization of Nonsingular Positive Rational Matrices, IEEE Transactions on Information Theory, vol. 21, no. 6, November 1975

    1975

  27. [35]

    Napp Avelli and H.L

    D. Napp Avelli and H.L. Trentelman, Algorithms for multidimensional spectral factorization and sum of squares, Linear Algebra and its Appli- cations, vol. 429, pp. 1114-1134, 2008

    2008

  28. [36]

    Pillai and J

    H.K. Pillai and J. C. Willems, Lossless and dissipative distributed sys- tems, SIAM Journal on Control & Optimization, vol. 40, pp. 1406-1430, 2002. 63

    2002

  29. [37]

    S. Y. Kung, B. C. Levy, Martin Morf, and Thomas Kailath, New results in 2-D system theory, Part II: 2-D State space model-realization and the notions of controllability, observability and minimality, Proceedings of the IEEE, vol. 65, no. 6, pp. 9, June 1977

    1977

  30. [38]

    M. P. Ekstrom , J. Woods, Two dimensional spectral factorization with applications in digital filtering, IEEE Transactions on ASSP, pp.115-128, vol. 24, April 1976

    1976

  31. [39]

    Parillo, Rekha R

    Grigoriy Blekherman, Pablo A. Parillo, Rekha R. Thomas (editors), Semidefinite Optimization and Convex Algebraic Geometry, MOS-SIAM Series on Optimization, vol. 13, December 2012

    2012

  32. [40]

    22, pp.27–44, 2011

    Joseph Ball, Multidimensional circuit synthesis and multivariable di- lation theory, Multidimensional Systems & Signal Processing, vol. 22, pp.27–44, 2011. 64

    2011

Pith tools

Reviewed July 12, 2026 · model on record in the stance chip above.