For a finitely generated extension G\mathcal G of a differential field F\mathcal F of characteristic zero with mm commuting derivations, the Kolchin differential dimension polynomial of a generating set depends on the chosen generators, and Sit's minimal differential dimension polynomial ωG/F\omega_{\mathcal G/\mathcal F}, defined as the minimum over all generating sets and guaranteed to exist by a well-ordering theorem, is the canonical generator-independent invariant of the extension. Two open questions arise from Kondratieva's study of the coefficients of these polynomials (2022/2023): (A) no effective, generally applicable procedure is known that, given generators (or a primitive single generator) of G/F\mathcal G/\mathcal F, computes ωG/F\omega_{\mathcal G/\mathcal F} with a guarantee of minimality, and finding a primitive element provably does not suffice; (B) over F=C(x1,x2)\mathcal F=\mathbb C(x_1,x_2) a family of singly generated extensions has generator polynomials with arbitrarily large Macaulay constants while the minimal polynomial stays fixed at 2s+12s+1, whereas for linear systems in one unknown with constant coefficients and a one-equation-shape minimal polynomial the source paper proves such a separation cannot occur for the generator the system itself defines, and the coefficient-field hypotheses that permit or prevent such unboundedness remain unclassified.

Background

A differential field (F,Δ)(\mathcal F,\Delta) is a field of characteristic zero equipped with finitely many commuting derivations Δ={∂1,…,∂m}\Delta=\{\partial_1,\dots,\partial_m\} (maps satisfying the Leibniz rule); the basic example is C(x1,…,xm)\mathbb C(x_1,\dots,x_m) with ∂i=∂/∂xi\partial_i=\partial/\partial x_i. A finitely generated differential (Δ\Delta-) extension G=F⟨ψ1,…,ψn⟩\mathcal G=\mathcal F\langle\psi_1,\dots,\psi_n\rangle is a differential field obtained from F\mathcal F by adjoining finitely many elements together with all their iterated derivatives. Kolchin showed that the transcendence degree over F\mathcal F of the field generated by all derivatives of the ψj\psi_j of total order at most ss is, for all sufficiently large ss, a numerical polynomial (a polynomial taking integer values at integers), namely

ωψ1,…,ψn/F(s)=trdeg⁡FF(θψj: θ∈Θ, ord⁡θ≤s, 1≤j≤n),\omega_{\psi_1,\dots,\psi_n/\mathcal F}(s)=\operatorname{trdeg}_{\mathcal F}\mathcal F\bigl(\theta\psi_j:\ \theta\in\Theta,\ \operatorname{ord}\theta\le s,\ 1\le j\le n\bigr),

where Θ\Theta is the free commutative monoid of derivative operators and ord⁡θ\operatorname{ord}\theta is the total order of θ\theta. This Kolchin differential dimension polynomial plays for differential algebraic varieties the role the Hilbert polynomial plays in commutative algebra: its degree (the differential type) and its leading coefficient (the typical differential dimension) are invariants of the extension. Unlike the Hilbert polynomial, however, it genuinely depends on the chosen generating set; for instance, adjoining one full round of derivatives of all generators shifts the argument by one.

Two encodings of the coefficient data of a numerical polynomial are central. The standard coefficients (ad,…,a0)(a_d,\dots,a_0) are defined by the expansion ω(s)=∑i=0dai(s+ii)\omega(s)=\sum_{i=0}^d a_i\binom{s+i}{i}. Sit's minimizing coefficients b(ω)=(bd,…,b0)b(\omega)=(b_d,\dots,b_0) are defined by a recurrence that strips off the leading standard coefficient, and they characterize the Kolchin polynomials among all numerical polynomials: a numerical polynomial is a Kolchin dimension polynomial exactly when all its minimizing coefficients are non-negative. Macaulay constants (cd+1,…,c1)(c_{d+1},\dots,c_1) of a numerical polynomial originate in Macaulay's theorem on Hilbert functions and appear, for example, in Dubé's degree bounds for Gröbner bases (special generating sets of polynomial ideals adapted to effective computation); they are read off from the unique representation ω(s)=(s+d+1d+1)−∑i=0d+1(s+i−1−cii)\omega(s)=\binom{s+d+1}{d+1}-\sum_{i=0}^{d+1}\binom{s+i-1-c_i}{i}. The source paper (Kondratieva 2022, published 2023) proves the closed formula expressing the Macaulay constants through the minimizing coefficients and deduces that the Macaulay constants of a Kolchin polynomial are non-decreasing.

Because the Kolchin polynomial depends on generators, Sit defined the minimal differential dimension polynomial: ωG/F\omega_{\mathcal G/\mathcal F} is the polynomial of some generating set that is eventually dominated by the polynomial of every generating set of G\mathcal G over F\mathcal F. Existence follows from Sit's theorem that the relevant numerical polynomials are well-ordered by eventual domination (equivalently, by the lexicographic order on minimizing coefficients read from the top). Kondratieva described in 1988 which numerical polynomials arise as minimal polynomials of some extension — an existence-level description of the value set, not a per-extension algorithm. For a fixed presentation, computing ωψ1,…,ψn/F\omega_{\psi_1,\dots,\psi_n/\mathcal F} and searching for a primitive element (a single element generating the extension, which exists under mild hypotheses, for instance when F⊇C(x1,…,xm)\mathcal F\supseteq\mathbb C(x_1,\dots,x_m) and the polynomial degree is below mm, by a theorem of Kolchin) is, in the source paper's words, only "in principle" solved, via the Rosenfeld–Gröbner algorithm — a differential analogue of Gröbner-basis decomposition, implemented in the Maple package diffag. Crucially, a primitive generator's polynomial need not be minimal: Example 5.7.7 of the Kondratieva–Levin–Mikhalev–Pankratiev monograph is the documented counterexample, and it is the reason the primitive-element construction does not solve the minimality problem. Adjacent algorithmic work stays on the fixed-presentation side: Lange-Hegermann computes the differential dimension polynomial for characterizable differential ideals (ideals encoded by differential Gröbner-basis-like data), and León Sánchez bounds all coefficients of the Kolchin polynomial of every prime component of a system in terms of the order of that system, answering a long-standing question of Kolchin. On the effectiveness frontier, the generalized Ritt problem — deciding inclusion of prime differential ideals given by their characteristic sets (Ritt's canonical encodings of prime differential ideals) — remains open (later work of Freitag, León Sánchez and Li confirms that the definability of full irreducibility for differential varieties is equivalent to it), delimiting what is currently computable in this area.

On the coefficient side, the source paper contrasts two regimes for singly generated extensions. Over the non-constant coefficient field F=C(x1,x2)\mathcal F=\mathbb C(x_1,x_2), its Example 2 exhibits, for every integer k≥2k\ge 2, a linear differential system in one unknown whose solution field G\mathcal G has generator polynomial ωφ/F=2s+k\omega_{\varphi/\mathcal F}=2s+k with minimizing coefficients (2,k−1)(2,k-1), while the explicit invertible change of generator ψ=φ−x1x2(∂1k−1φ−x2∂1k−1∂2φ)\psi=\varphi-x_1x_2(\partial_1^{k-1}\varphi-x_2\partial_1^{k-1}\partial_2\varphi) satisfies the constant-coefficient equation ∂1∂2ψ=0\partial_1\partial_2\psi=0 with ωψ/F=ωG/F=2s+1\omega_{\psi/\mathcal F}=\omega_{\mathcal G/\mathcal F}=2s+1; hence the Macaulay constants of a generator's polynomial are unbounded in terms of the minimal polynomial alone. Note that the field of Example 2 is itself generated by an element satisfying a constant-coefficient equation, whose Kolchin polynomial equals the minimal polynomial; the unbounded generators φ\varphi arise from non-constant-coefficient presentations of the same field. Consequently, over all generators ξ\xi of G\mathcal G the Macaulay constants are unbounded in terms of ωG/F\omega_{\mathcal G/\mathcal F} in either coefficient regime, so a boundedness question is meaningful only for the generator defined by the given system. By contrast, the paper's Theorem 5 shows that for linear systems in one unknown whose coefficients are constants of F\mathcal F, the Example-2 separation cannot occur for the generator such a system defines whenever the minimal polynomial has the one-equation shape: if the minimal polynomial of G/F\mathcal G/\mathcal F is (s+mm)−(s+m−dm)\binom{s+m}{m}-\binom{s+m-d}{m}, whose Macaulay constants are all equal, then Theorem 5 forces the system generator's Kolchin polynomial to have that same shape, and Theorem 4 then forces it to equal the minimal polynomial itself; the theorem is silent for minimal polynomials of other shapes. The proof reduces the situation to the commutative polynomial ring of differential operators with constant coefficients (the skew, Ore-type structure of F[∂1,…,∂m]\mathcal F[\partial_1,\dots,\partial_m], in which derivations act non-trivially on non-constant coefficients, being trivial on the constants) and shows that the operator ideal attached to the system is projective and therefore principal. The paper's final remark identifies the Example-2 construction with a projective non-free left ideal of the noncommutative ring C(x1,x2)[∂1,∂2]\mathbb C(x_1,x_2)[\partial_1,\partial_2], pinpointing the module-theoretic mechanism of the blow-up.

Problem Statement

The problem concerns the minimal differential dimension polynomial ωG/F\omega_{\mathcal G/\mathcal F} of a finitely generated differential extension G=F⟨ψ1,…,ψn⟩\mathcal G=\mathcal F\langle\psi_1,\dots,\psi_n\rangle of a differential field of characteristic zero with mm commuting derivations, and has two linked components.

(A) Effective determination. Given an effective presentation of F\mathcal F and of G\mathcal G over F\mathcal F — the generators specified by a finite system of differential polynomial equations, with inequations where necessary to isolate a prime differential component — construct a generally applicable, terminating algorithm that outputs ωG/F\omega_{\mathcal G/\mathcal F} together with a proof of minimality: a proof that the output equals ωη0/F\omega_{\eta_0/\mathcal F} for some generating set η0\eta_0 of G\mathcal G and is eventually dominated by ωη/F\omega_{\eta/\mathcal F} for every generating set η\eta of G\mathcal G. The known ingredients — computing ωψ1,…,ψn/F\omega_{\psi_1,\dots,\psi_n/\mathcal F} for the given generators and searching for a primitive element via Rosenfeld–Gröbner-type decompositions — do not constitute a solution, since a primitive generator's polynomial need not be minimal.

(B) Coefficient-field classification of boundedness. Characterize the hypotheses on the base field F\mathcal F and on the coefficients of the defining system under which the Macaulay constants (equivalently, the minimizing coefficients) of the Kolchin polynomial ωξ/F\omega_{\xi/\mathcal F} of the generator ξ\xi defined by such a system are bounded by a quantity depending only on the minimal polynomial ωG/F\omega_{\mathcal G/\mathcal F} and the number of derivations mm, as opposed to unbounded. The restriction to the generator defined by the given system is essential to the question: as the source paper's Example 2 shows, the field generated by an element ψ\psi with the constant-coefficient equation ∂1∂2ψ=0\partial_1\partial_2\psi=0 and minimal polynomial 2s+12s+1 also admits invertible changes of generator φ\varphi with ωφ/F=2s+k\omega_{\varphi/\mathcal F}=2s+k for every k≥2k\ge 2, so quantification over all generators of G\mathcal G would make unboundedness hold trivially in every coefficient regime. The two established regimes are: coefficients of the defining linear system that are constants of F\mathcal F, where the source paper's Theorem 5 proves that a minimal polynomial of the one-equation shape (s+mm)−(s+m−dm)\binom{s+m}{m}-\binom{s+m-d}{m} forces the system generator's Kolchin polynomial to equal it, so that the Example-2 blow-up cannot occur for that generator (the theorem does not cover minimal polynomials of other shapes); and F=C(x1,x2)\mathcal F=\mathbb C(x_1,x_2) with non-constant coefficients, where the Example 2 family realizes arbitrarily large Macaulay constants against the fixed minimal polynomial 2s+12s+1. Which coefficient-field conditions separate these two behaviours is the open content of the classification.

The common objective is effective and structural control of the minimal differential dimension polynomial: component (A) asks for computation of the invariant itself, and component (B) delimits the coefficient behaviour of generating systems that any such computation must confront.

References

  1. M. V. Kondratieva, Some Properties of Coefficients Kolchin Dimension Polynomial, arXiv:2202.02542 (2022); published as Some Properties of Coefficients of the Kolchin Dimension Polynomial, Journal of Mathematical Sciences 269 (2023) 725-733, DOI 10.1007/s10958-023-06309-0, https://arxiv.org/abs/2202.02542
  2. E. R. Kolchin, The notion of dimension in the theory of algebraic differential equations, Bulletin of the American Mathematical Society 70 (1964) 570-573, DOI 10.1090/S0002-9904-1964-11203-9, https://doi.org/10.1090/S0002-9904-1964-11203-9
  3. E. R. Kolchin, Differential Algebra and Algebraic Groups, Pure and Applied Mathematics 54, Academic Press, New York (1973), zbMATH 3415973, https://zbmath.org/?q=an:3415973
  4. William Yu. Sit, Well-ordering of certain numerical polynomials, Transactions of the American Mathematical Society 212 (1975) 37-45, DOI 10.1090/S0002-9947-1975-0406992-9, https://doi.org/10.1090/S0002-9947-1975-0406992-9
  5. William Sit, Differential dimension polynomials of finitely generated extensions, Proceedings of the American Mathematical Society 68 (1978) 251-257, DOI 10.1090/S0002-9939-1978-0480353-7, https://doi.org/10.1090/S0002-9939-1978-0480353-7
  6. M. V. Kondratieva, A. B. Levin, A. V. Mikhalev, E. V. Pankratiev, Differential and Difference Dimension Polynomials, Kluwer Academic Publishers (1999), DOI 10.1007/978-94-017-1257-6, https://doi.org/10.1007/978-94-017-1257-6
  7. M. V. Kondrat'eva, Description of the set of minimal differential dimensional polynomials, Vestnik Moskovskogo Universiteta, Seriya I: Matematika, Mekhanika, 1988, no. 1, 35-39 (Russian; English translation: A description of a set of minimal differential dimensional polynomials, Moscow University Mathematics Bulletin 43 (1988), no. 1, 46-50, zbMATH 4065127), zbMATH 4037960, https://zbmath.org/?q=an:4037960
  8. M. V. Kondratieva, V. A. Mityunin, Computation of the Differential Dimension Polynomial When Changing Generators in the Dirac Equations, Programming and Computer Software 27 (2001) 25-28, DOI 10.1023/A:1007182602446, https://doi.org/10.1023/A:1007182602446
  9. M. V. Kondratieva, An upper bound for minimizing coefficients of dimension Kolchin polynomial, Programming and Computer Software 36 (2010) 83-86, DOI 10.1134/S0361768810020052, https://doi.org/10.1134/S0361768810020052
  10. O. D. Golubitsky, M. V. Kondratieva, A. I. Ovchinnikov, On the generalized Ritt problem as a computational problem, Journal of Mathematical Sciences 163 (2009) 515-522, DOI 10.1007/s10958-009-9689-3; preprint arXiv:0809.1128 (preprint title: On the generalised Ritt problem as a computational problem), https://arxiv.org/abs/0809.1128
  11. Markus Lange-Hegermann, The Differential Dimension Polynomial for Characterizable Differential Ideals, in Algorithmic and Experimental Methods in Algebra, Geometry, and Number Theory, Springer (2017) 443-453, DOI 10.1007/978-3-319-70566-8_18; preprint arXiv:1401.5959, https://arxiv.org/abs/1401.5959
  12. Omar León Sánchez, Estimates for the coefficients of differential dimension polynomials, Mathematics of Computation 88 (2019) 2959-2985, DOI 10.1090/mcom/3429; preprint arXiv:1703.00509, https://arxiv.org/abs/1703.00509
  13. James Freitag, Omar León Sánchez, Wei Li, Effective definability of Kolchin polynomials, Proceedings of the American Mathematical Society 148 (2020) 1455-1466, DOI 10.1090/proc/14869; preprint arXiv:1806.02060, https://arxiv.org/abs/1806.02060
  14. M. V. Kondratieva, A Bound for a Typical Differential Dimension of Systems of Linear Differential Equations, Journal of Mathematical Sciences 259 (2021) 563-569, DOI 10.1007/s10958-021-05647-1; preprint arXiv:1802.06163 (preprint title: A bound for a typical differential dimension of system of linear differential equations), https://arxiv.org/abs/1802.06163
  15. M. V. Kondratieva, The typical dimension of a system of first-order differential equations, Journal of Mathematical Sciences 283 (2024) 912-918, DOI 10.1007/s10958-024-07319-2, https://doi.org/10.1007/s10958-024-07319-2
  16. Alexander Levin, Bivariate Differential Dimension Polynomials via Bigraded Rees Construction, Proceedings of the 2026 International Symposium on Symbolic and Algebraic Computation (ISSAC), ACM (2026) 264-271, DOI 10.1145/3815436.3815456, https://doi.org/10.1145/3815436.3815456
  17. Thomas W. Dubé, The Structure of Polynomial Ideals and Gröbner Bases, SIAM Journal on Computing 19 (1990) 750-773, DOI 10.1137/0219053, https://doi.org/10.1137/0219053