Estimates of the asymptotic Nikolskii constants for spherical polynomials
Pith reviewed 2026-05-25 00:31 UTC · model grok-4.3
The pith
The asymptotic Nikolskii constant L^*(d) for spherical polynomials satisfies 0.5^d ≤ L^*(d) ≤ (0.857⋯)^{d(1+ε_d)} with ε_d=O(d^{-2/3}).
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The paper establishes a close connection between L^*(d) and the extremal problem I_alpha, which is the infimum over coefficient sequences of the L^∞ norm of j_{alpha+1}(t) minus a sum of scaled j_alpha functions evaluated at the zeros of j_{alpha+1}. It proves that for alpha ≥ -0.272 this infimum equals the integral from 0 to q_{alpha+1,1} of j_{alpha+1}(t) t^{2alpha+1} dt divided by the integral from 0 to q_{alpha+1,1} of t^{2alpha+1} dt, and that this common value is the hypergeometric _1F_2(alpha+1; alpha+2, alpha+2; -q_{alpha+1,1}^2/4). As a direct result the bounds 0.5^d ≤ L^*(d) ≤ (0.857⋯)^{d(1+ε_d)} hold with ε_d = O(d^{-2/3}).
What carries the argument
The extremal problem I_alpha, defined as the infimum L^∞ distance from j_{alpha+1} to absolutely convergent linear combinations of scaled copies of j_alpha at the positive zeros of j_{alpha+1}.
If this is right
- L^*(d) tends to zero exponentially fast with growing dimension d.
- The upper bound includes a sub-exponential correction factor that still tends to 1.
- The lower bound of exactly 0.5^d is obtained from the same connection.
- The hypergeometric representation supplies an explicit closed form for I_alpha in the stated range of alpha.
Where Pith is reading between the lines
- The exponential decay implies that in high dimensions almost all mass of a typical spherical polynomial is concentrated near its average value, which may affect discrepancy or quadrature error estimates.
- The integral representation of I_alpha could be used to obtain sharper asymptotics for alpha near the boundary value -0.272.
- Similar connections between Nikolskii-type constants and Bessel extremal problems may exist on other manifolds or with different measures.
- Testing the equality for I_alpha at alpha values slightly below -0.272 would show whether the threshold is sharp.
- keywords:[
Load-bearing premise
The abstract connection between the Nikolskii constant L^*(d) and the Bessel extremal value I_alpha remains valid when alpha is linked to dimension d and is at least -0.272.
What would settle it
Direct numerical evaluation of sup ||f||_∞ / ||f||_1 over a fine net of spherical polynomials of increasing degree in dimensions d=10,20,30 would lie outside the stated exponential envelopes if the bounds are incorrect.
read the original abstract
Let $\Pi_n^d$ denote the space of spherical polynomials of degree at most $n$ on the unit sphere $\mathbb{S}^d\subset \mathbb{R}^{d+1}$ that is equipped with the surface Lebesgue measure $d\sigma$ normalized by $\int_{\mathbb{S}^d} \, d\sigma(x)=1$. This paper establishes a close connection between the asymptotic Nikolskii constant, $$ \mathcal{L}^\ast(d):=\lim_{n\to \infty} \frac 1 {\dim \Pi_n^d} \sup_{f\in \Pi_n^d} \frac { \|f\|_{L^\infty(\mathbb{S}^d)}}{\|f\|_{L^1(\mathbb{S}^d)}},$$ and the following extremal problem: $$ \mathcal{I}_\alpha:=\inf_{a_k} \Bigl\| j_{\alpha+1} (t)- \sum_{k=1}^\infty a_k j_{\alpha} \bigl( q_{\alpha+1,k}t/q_{\alpha+1,1}\bigr)\Bigr\|_{L^\infty(\mathbb{R}_+)} $$ with the infimum being taken over all sequences $\{a_k\}_{k=1}^\infty\subset \mathbb{R}$ such that the infinite series converges absolutely a.e. on $\mathbb{R}_+$. Here $j_\alpha $ denotes the Bessel function of the first kind normalized so that $j_\alpha(0)=1$, and $\{q_{\alpha+1,k}\}_{k=1}^\infty$ denotes the strict increasing sequence of all positive zeros of $j_{\alpha+1}$. We prove that for $\alpha\ge -0.272$, $$\mathcal{I}_\alpha= \frac{\int_{0}^{q_{\alpha+1,1}}j_{\alpha+1}(t)t^{2\alpha+1}\,dt}{\int_{0}^{q_{\alpha+1,1}}t^{2\alpha+1}\,dt}= {}_{1}F_{2}\Bigl(\alpha+1;\alpha+2,\alpha+2;-\frac{q_{\alpha+1,1}^{2}}{4}\Bigr). $$ As a result, we deduce that the constant $\mathcal{L}^\ast(d)$ goes to zero exponentially fast as $d\to\infty$: \[ 0.5^d\le \mathcal{L}^{*}(d)\le (0.857\cdots)^{d\,(1+\varepsilon_d)} \ \ \ \ \ \text{with $\varepsilon_d =O(d^{-2/3})$.} \]
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper connects the asymptotic Nikolskii constant L^*(d) for spherical polynomials on the d-dimensional sphere to an extremal approximation problem I_α involving normalized Bessel functions j_α. It proves that for α ≥ -0.272, I_α equals the ratio of the integral from 0 to q_{α+1,1} of j_{α+1}(t) t^{2α+1} dt to the integral of t^{2α+1} dt, which is also equal to the hypergeometric function _1F_2(α+1; α+2, α+2; -q_{α+1,1}^2/4). Using this, it deduces exponential bounds 0.5^d ≤ L^*(d) ≤ (0.857⋯)^{d(1+ε_d)} with ε_d = O(d^{-2/3}).
Significance. If the central equality holds, the paper provides explicit exponential decay rates for L^*(d) as d increases, which is significant for understanding the behavior of spherical polynomials in high dimensions. The link to Bessel function extremal problems and the closed-form expression via hypergeometric series strengthens the result. The lower bound is parameter-free and the upper bound gives a concrete base less than 1.
major comments (1)
- [Abstract] Abstract: The equality I_α = ∫_0^{q_{α+1,1}} j_{α+1}(t) t^{2α+1} dt / ∫_0^{q_{α+1,1}} t^{2α+1} dt (equivalently the _1F_2 expression) is asserted to hold for all α ≥ −0.272, but the threshold −0.272 is stated numerically to three decimals. The manuscript must clarify how this range is established rigorously, in particular whether the L^∞ tail control outside [0, q_{α+1,1}] for the approximating series is proven analytically for the entire interval or relies on numerical verification that the infimum is attained by the integral ratio.
minor comments (1)
- [Abstract] Abstract: The definition of the sequence {q_{α+1,k}} as the positive zeros of j_{α+1} is stated but could be cross-referenced to the standard notation for Bessel zeros used later in the extremal problem.
Simulated Author's Rebuttal
Thank you for your thorough review of our manuscript. We address the major comment below and will incorporate the necessary clarifications in the revised version.
read point-by-point responses
-
Referee: [Abstract] Abstract: The equality I_α = ∫_0^{q_{α+1,1}} j_{α+1}(t) t^{2α+1} dt / ∫_0^{q_{α+1,1}} t^{2α+1} dt (equivalently the _1F_2 expression) is asserted to hold for all α ≥ −0.272, but the threshold −0.272 is stated numerically to three decimals. The manuscript must clarify how this range is established rigorously, in particular whether the L^∞ tail control outside [0, q_{α+1,1}] for the approximating series is proven analytically for the entire interval or relies on numerical verification that the infimum is attained by the integral ratio.
Authors: We thank the referee for pointing out the need for clarification on this point. The proof that I_α equals the given integral ratio (and thus the hypergeometric expression) is analytical and holds whenever the L^∞ norm of the remainder term outside [0, q_{α+1,1}] is bounded by the value of the integral ratio itself. This condition on the tail control is verified numerically for α ≥ −0.272, and we have computed that −0.272 is approximately the infimum of such α where this holds. In the revised manuscript, we will add a detailed explanation of this analytical condition and describe the numerical procedure used to determine the threshold, including the precision of the computation, to make the range rigorous. revision: yes
Circularity Check
No significant circularity; the equality I_α = integral ratio is a proved result for α ≥ −0.272, not a definitional reduction.
full rationale
The paper defines I_α explicitly as an infimum over coefficient sequences {a_k} of the L^∞(R_+) distance to j_{α+1}(t). It then proves (not assumes) that this infimum equals the weighted-integral ratio on [0, q_{α+1,1}] for the stated range of α, with the ratio also equal to the _1F_2 series. The lower bound on the infimum follows from the standard annihilator argument using the weight t^{2α+1} (which kills the approximants by orthogonality properties of the scaled Bessel functions), while equality requires an explicit construction or existence proof that the error does not exceed this value outside the interval; both directions are independent mathematical content. The link from I_α to the asymptotic Nikolskii constant L^*(d) is likewise established inside the paper rather than imported by self-citation or ansatz. No step reduces by construction to its own inputs, and the exponential bounds on L^*(d) are consequences of this analysis rather than tautological renamings or fitted predictions.
Axiom & Free-Parameter Ledger
free parameters (1)
- alpha threshold =
-0.272
axioms (1)
- standard math Standard analytic properties of the normalized Bessel function j_α and the sequence of its positive zeros q_{α+1,k}
Reference graph
Works this paper leans on
-
[1]
D. Amir and Z. Ziegler, Polynomials of extremal Lp-norm on the L∞-unit sphere, J. Approx. Theory 18 (1976), 86–98
work page 1976
-
[2]
N. N. Andreev, S. V. Konyagin, and A. Yu. Popov, Extremum problems for functions with small support, Math. Notes 60 (1996), no. 3, 241–247; Math. Notes 68 (2000), no. 3, 415
work page 1996
-
[3]
V. V. Arestov, Inequality of different metrics for trigonometric polynomi als, Math. Notes 27 (1980), no. 4, 265–269
work page 1980
-
[4]
V. V. Arestov and M. V. Deikalova, Nikol’skii inequality for algebraic polynomials on a multi dimen- sional Euclidean sphere , Proc. Steklov Inst. Math. 284 (2014), Suppl. 1, S9–S23
work page 2014
-
[5]
V. Arestov, A. Babenko, M. Deikalova, and ´A. Horv´ ath,Nikol’skii inequality between the uniform norm and integral norm with Bessel weight for entire functio ns of exponential type on the half-line , Anal. Math., 44 (2018), no. 1, 21–42
work page 2018
-
[6]
G. Bateman, A. Erd´ elyi, et al., Higher Transcendental Functions, Vol. II, McGraw Hill Book Com- pany, New York, 1953
work page 1953
- [7]
-
[8]
Y. Benyamini, A. Kro´ o, and A. Pinkus, L1-approximation and finding solutions with small support , Constr. Approx. 36 (2012), no. 3, 399–431
work page 2012
-
[9]
R. P. Boas, Entire Functions, N. Y.: Academic Press, 1954
work page 1954
-
[10]
E. Carneiro, M. B. Milinovich, and K. Soundararajan, Fourier optimization and prime gaps , arXiv:1708.04122
-
[11]
H. Cohn, A. Kumar, S. D. Miller, D. Radchenko, and M. Viazovska , The sphere packing problem in dimension 24, Ann. of Math., 185 (2017), 1017–1033
work page 2017
-
[12]
H. Cohn and N. Elkies, New upper bounds on sphere packings. I , Ann. of Math. (2) 157 (2003), no. 2, 689–714
work page 2003
- [13]
-
[14]
F. Dai, D. Gorbachev, and S. Tikhonov, Nikolskii constants for polynomials on the unit sphere , J. d’Analyse Math. (2019) (in press); arXiv:1708.09837
work page internal anchor Pith review Pith/arXiv arXiv 2019
-
[15]
R. DeVore, and G. G. Lorentz, Constructive approximation. G rundlehren der Mathematischen Wis- senschaften, 303. Springer-Verlag, Berlin, 1993
work page 1993
-
[16]
Z. Ditzian, S. Tikhonov, Ul’yanov and Nikol’skii-type inequalities , J. Approx. Theory, 133 (2005), no. 1, 100–133
work page 2005
-
[17]
M. Ganzburg and S. Tikhonov, On sharp constants in Bernstein–Nikolskii inequalities , Constr. Ap- prox. 45 (2017), no. 3, 449–466
work page 2017
-
[18]
Geronimus, Sur un probl` eme extr´ emal de Tchebycheff, Izv
J. Geronimus, Sur un probl` eme extr´ emal de Tchebycheff, Izv. Akad. Nauk SSSR Ser. Mat. 2 (1938), no. 4, 445–456
work page 1938
-
[19]
A. V. Gladkaya and O. L. Vinogradov, Entire functions that deviate least from zero in the uniform and the integral metrics with a weight , St. Petersburg Math. J. 26 (2015), 867–879
work page 2015
-
[20]
D. V. Gorbachev, Extremal problem for entire functions of exponential spher ical type, connected with the Levenshtein bound on the sphere packing density in Rn (in Russian), Izvestiya of the Tula State University Ser. Mathematics Mechanics Informatics, 6 (2000), 71–78
work page 2000
-
[21]
D. V. Gorbachev, Extremum problems for entire functions of exponential sphe rical type, Math. Notes 68 (2000), no. 2, 159–166
work page 2000
-
[22]
D. V. Gorbachev, An integral problem of Konyagin and the (C, L)-constants of Nikol’skii , Proc. Steklov Inst. Math. Suppl. 2 (2005), S117–S138
work page 2005
-
[23]
D. V. Gorbachev, V. I. Ivanov, On interrelation of Nikolskii constants for trigonometric polynomials and entire functions of exponential type , Chebyshevskii Sbornik, 19 (2018), no. 2, 80–89. (In Russ.) https://doi.org/10.22405/2226-8383-2018-19-2-80-89 26 FENG DAI, DMITRY GORBACHEV, AND SERGEY TIKHONOV
-
[24]
T. K. Ho, An inequality for algebraic polynomials, and the dependenc e between the best polynomial approximations E(f )Lp and E(f )Lq of functions f (x) ∈ Lp, Acta Math. Acad. Sci. Hung. 27 (1976), 141–147
work page 1976
-
[25]
L. H¨ ormander and B. Bernhardsson, An extension of Bohr’s inequality , Boundary value problems for partial differential equations and applications, RMA Res. Notes App l. Math., 29, 179–194 (1993)
work page 1993
-
[26]
Jackson, Certain problems of closest approximation , Bull
D. Jackson, Certain problems of closest approximation , Bull. Am. Math. Soc. 39 (1933), 889–906
work page 1933
-
[27]
A. I. Kamzolov, Approximation of functions on the sphere Sn, Serdica 84 (1984), no. 1, 3–10
work page 1984
-
[28]
V. Levenshtein, On designs in compact metric spaces and a universal bound on t heir size , Discrete Math. 192 (1998), 251–271
work page 1998
-
[29]
E. Levin and D. Lubinsky, Lp Chritoffel functions, Lp universality, and Paley–Wiener spaces , J. D’Analyse Math. 125 (2015), 243–283
work page 2015
-
[30]
E. Levin and D. Lubinsky, Asymptotic behavior of Nikolskii constants for polynomial s on the unit circle, Comput. Methods Funct. Theory 15 (2015), no. 3, 459–468
work page 2015
-
[31]
F. Littmann, M. Spanier, Extremal signatures , Constr Approx. (2017). https://doi.org/10.1007/ s00365-017-9373-7; arXiv:1609.03987
work page internal anchor Pith review Pith/arXiv arXiv 2017
-
[32]
Y. L. Luke, The Special Functions and their Approximations , Vol. 1. Academic Press, New York, 1969
work page 1969
-
[33]
Yu. V. Malykhin, K. S. Ryutin, Concentration of the L1-norm of trigonometric polynomials and entire functions , Sb. Math., 205 (2014), no. 11, 1620–1649
work page 2014
-
[34]
Nemes, On the large argument asymptotics of the Lommel function via Stieltjes transforms , Asymptot
G. Nemes, On the large argument asymptotics of the Lommel function via Stieltjes transforms , Asymptot. Anal. 91 (2015), no. 3-4, 265–281
work page 2015
-
[35]
R. Nessel and G. Wilmes, Nikolskii-type inequalities for trigonometric polynomia ls and entire func- tions of exponential type , J. Austral. Math. Soc., 25 (1978), no. 1, 7–18
work page 1978
-
[36]
S. M. Nikolskii, Approximation of functions of several variables and imbedd ing theorems , Berlin; Heidelberg; New York: Springer, 1975
work page 1975
-
[37]
F. W. J. Olver, D. W. Lozier, R. F. Boisvert, and C. W. Clark, eds . NIST Handbook of Mathematical Functions, Cambridge University Press, New York, 2010; NIST Digital Library of Mathematical Functions. http://dlmf.nist.gov/, Release 1.0.18 of 2018-03-27
work page 2010
-
[38]
C. K. Qu and R. Wong, “Best possible” upper and lower bounds for the zeros of the Be ssel function Jν(x), Trans. Amer. Math. Soc. 351 (1999), no. 7, 2833–2859
work page 1999
-
[39]
Szeg¨ o,Orthogonal Polynomials, Amer
G. Szeg¨ o,Orthogonal Polynomials, Amer. Math. Soc., New York, 1967
work page 1967
-
[40]
Shapiro, Topics in approximation theory , Lecture Notes in Mathematics, Vol
H. Shapiro, Topics in approximation theory , Lecture Notes in Mathematics, Vol. 187, Springer- Verlag, Berlin, Heidelberg, 1971
work page 1971
-
[41]
L. V. Taikov, A group of extremal problems for trigonometric polynomials , Uspekhi Mat. Nauk (in Russian), 20 (1965), no. 3(123), 205—211
work page 1965
-
[42]
L. V. Taikov, On the best approximation of Dirichlet kernels , Math. Notes 53 (1993), no. 6, 640–643
work page 1993
-
[43]
V. Temlyakov, S. Tikhonov. Remez-type and Nikol’skii-type inequalities: general rel ations and the hyperbolic cross polynomials, Constr. Approx. 46 (2017), no. 3, 593–615
work page 2017
-
[44]
M. S. Viazovska, The sphere packing problem in dimension 8, Ann. of Math. 185 (2017), no. 3, 991–1015
work page 2017
-
[45]
G. N. Watson, A Treatise on the Theory of Bessel Functions (2nd ed.) Cambridge University Press, 1966. 27 F. Dai, Department of Mathematical and Statistical Science s, University of Alberta, Edmonton, Alberta T6G 2G1, Canada. E-mail address : fdai@ualberta.ca D. Gorbachev, Tula State University, Department of Applied Mathematics and Com- puter Science, 30...
work page 1966
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.