Regularized Bessel moments

edit · history · discussion · files · short url · integral moments period special values number theory
Numbers
$n$
$j$
$\mathrm{normalization}$ 
$F_{n,j}$ (bessel) or $G_{n,j}$ (period)
4
1
bessel:
0.01983571147245574644959672583808746210113166219478452818373555495073732206452230558523848887798341698
4
1
period:
-0.7830825009891516688698323331023089496753115878661001127917353846654119718777778506214116446554323431
8
2
bessel:
0.02016227366182216956447644863453487769960662960981970832235331891286295440137676027634710691114950338
8
2
period:
31.42382000923048617045694512927379118867452236612118140035451872667564235798830700646753174311839636
8
3
bessel:
0.007026505713390779904991693926620146478765556002286598375865619120685230536400308659572663392570030499
8
3
period:
2.737782138746482871349466666223353898237583613936255207686684850595939140612943660285815830396158272
12
3
bessel:
0.009770700556352949550349321728813232848027621930176655658845099438035651754981660955921894363745011677
12
3
period:
-601.1805394264281967579861013901833076268444819684588826042858454578290294555897393052666796215903108
12
4
bessel:
0.004456721171628450865378663990468082063638025811732252907977755498438982815607834660983599347552112069
12
4
period:
-68.55429717090979740721458153317883120711971092025706673693120095025116084359392668121192462608525022
12
5
bessel:
0.004649372233084364227844917555920825066083710342478756690973194828491835083090069578012042997500369800
12
5
period:
-17.87942488718555720409771118423333208430553986708215556937536305031160368791205162255671978274660413
16
4
bessel:
0.003790331876250755729868960031389489139816053245700672499609743391961035367504947300175221866181482275
16
4
period:
9206.958481665604721959816364313953625469708552388081962244480085741476535963050428478188223408486136
16
5
bessel:
0.002115830257847242592603696761947495844527961991653360137741294856227523799784060195858133675432896599
16
5
period:
1284.871745684746326380582141903525488049800904345513585973141482999845675348293638813811444043414469
16
6
bessel:
0.002351229614266707451125066264032052911470379839037748637829720043704078907790503374213524151269468350
16
6
period:
356.9554419339728870681067485648141493458219118018460815064776377444116884629491622755394496401608018
16
7
bessel:
0.004821070999265147795852751087493435500260854010259714001385413878884491067784282842120255761351018945
16
7
period:
182.9795268288228414713694881244422014425681854785235586063127472222206695900472824344841329539370325
Definition
For $n=4m\geq4$ and $j\geq m$, the balanced Bessel finite part in (6) has two invertible conventions: bessel returns $F_{n,j}$ and period returns $G_{n,j}$, related by (1) [1].
Parameters
$n$
—   total number of Bessel factors ($n\geq4$, $4\mid n$)
$j$
—   half the integration power ($j\geq n/4$)
$\mathrm{normalization}$
—   moment normalization ($\mathrm{bessel}$ or $\mathrm{period}$)
Formulas
(1)
$G_{n,j}=(-1)^m2^{n-2j}\pi^{2m}F_{n,j}$, where $n=4m$ and $j\geq m$. This nonzero factor gives the inverse conversion as well.
(2)
For $n=4m$ and $m\leq j\leq2m-1$, $F_{4m,j}=(-1)^{m-1}\pi^{-2m}\sum_{a=0}^{m-1}(-1)^a\binom{2m}{2a}\pi^{2a}B_{4m}(2a,2j)$ [1], Corollary 28. Here $B_n(a,q)$ is the convergent moment defined in (9); every constituent has exponential decay at infinity.
(3)
At four factors, $F_{4,1}=B_4(0,2)/\pi^2$ and $G_{4,1}=-4B_4(0,2)$, with $B_4(0,2)$ the ordinary four-$K_0$ moment. These are the $m=1$ cases of the sum rule and the period conversion.
Comments
(4)
The functions $I_0$ and $K_0$ are the standard order-zero modified Bessel function of the first kind and modified Bessel function of the second kind, at common unit scale on the positive real axis. The positive-real principal branch of $K_0$ is used. No loop factors or factorials are absorbed.
(5)
Write $\gamma_{n,\ell}=[w^\ell](\sum_{s\geq0}((2s-1)!!)^3w^s/(2^{5s}s!))^{n/2}$, where $[w^\ell]$ means formal coefficient extraction, $!!$ denotes the double factorial and $(-1)!!=1$. This is a formal series, not a convergent expansion at nonzero $w$; in particular $\gamma_{n,0}=1$. The coefficients follow from the product expansion in [1], equation (10).
(6)
For $n=4m$, define $S_{n,j}(T)=2^{-2m}\sum_{\ell=0}^{j-m}4^\ell\gamma_{n,\ell}T^{2j-2m-2\ell+1}/(2j-2m-2\ell+1)$ and $F_{n,j}=\lim_{T\to\infty}(\int_0^T(I_0(t)K_0(t))^{2m}t^{2j}\,dt-S_{n,j}(T))$ [1], Lemma 21. Subtract exactly these nonintegrable powers and no constant. There is no logarithmic counterterm or reference scale in this branch.
(7)
The integrand is $O(t^{2j}(1+|\log t|)^{2m})$ at zero, where it is integrable. At infinity it has the asymptotic expansion $2^{-2m}\sum_{\ell\geq0}4^\ell\gamma_{n,\ell}t^{2j-2m-2\ell}$. Subtraction of $S_{n,j}^{\prime}(t)$ leaves $O(t^{-2})$, so the specified finite part exists [1], Lemma 21. The unsubtracted integral diverges for every allowed $j$.
(8)
The period convention is the compactly supported pairing $\langle\widetilde\omega_{n,j},\gamma_{n/2}\rangle_{\mathrm{per,c}}$ of [1], Proposition 25, on its period-basis slice $m\leq j\leq2m-1$, with $t=2\sqrt z$. The displayed formula for $G_{n,j}$ fixes the sign and normalization explicitly and defines its extension to all $j\geq m$. The number $\pi$ has its usual positive value. The two conventions determine one another by a nonzero elementary factor.
(9)
For convergent integrals write $B_n(a,q)=\int_0^\infty I_0(t)^aK_0(t)^{n-a}t^q\,dt$. This is $\operatorname{IKM}(a,n-a;q)$, while [1] write $\operatorname{IKM}_n(a,q)$. Ordinary convergent moments and divergent integrals with a prescribed finite part have different definitions.
Programs
(P1)
Python
from mpmath import mp

def regularized_bessel_moment(n, j, normalization='bessel', dps=120):
    # Corollary 28 on the divergent compactly supported period-basis slice.
    n, j = int(n), int(j)
    if n < 4 or n % 4 or not n//4 <= j <= n//2-1:
        raise ValueError('n=4m and m<=j<=2m-1')
    if normalization not in ('bessel', 'period'):
        raise ValueError('normalization bessel or period')
    with mp.workdps(dps):
        m = n//4
        def moment(a):
            def f(t):
                if not t: return mp.mpf(0)
                return mp.besseli(0,t)**a * mp.besselk(0,t)**(n-a) * t**(2*j)
            return mp.quad(f, [0, 1, 4, 16, mp.inf])
        F = (-1)**(m-1)/mp.pi**(2*m) * sum(
            (-1)**a * mp.binomial(2*m,2*a) * mp.pi**(2*a) * moment(2*a)
            for a in range(m))
        if normalization == 'period':
            F *= (-1)**m * mp.mpf(2)**(n-2*j) * mp.pi**(2*m)
        return +F

# A higher-factor example beyond the initial n<=16 selection:
# regularized_bessel_moment(20, 5, 'bessel')
# This invocation is numerical; it supplies no certified error bound.
References
[1]
Ping-Hsun Chuang and Jeng-Daw Yu, On the Periods of Twisted Moments of the Kloosterman Connection, The Ramanujan Journal 65 (2024), 1227–1277. ArXiv v2 (24 September 2024), equation (10), Lemma 21, Proposition 25 and Corollary 28. (arXiv) (doi)
Links
Similar tables
Bessel moments —   Ordinary convergent scalar moments entering the sum rule; the balanced integrals here require explicit power subtraction.
Determinants of even-power Bessel moments —   Full rapid-decay period determinants with consecutive first-kind exponents and even integration powers, rather than individual compactly supported balanced finite parts.
Bessel period determinants from odd rows —   Untwisted critical-period minors with odd first-kind row exponents and odd integration powers.
Bessel period determinants from even rows —   Untwisted critical-period minors with even first-kind row exponents and odd integration powers.
Broadhurst–Mellit determinants of Bessel moments —   Full determinants of ordinary moments with odd integration powers.
Broadhurst–Mellit vacuum determinants —   Three prescribed vacuum determinant constructions from ordinary moments.
Data properties
Entries are of type: real number
digits: 100 (Unknown key)
Table is complete: no (it holds both bessel and period for every $n=4m$, $1\leq m\leq4$, and $m\leq j\leq2m-1$. These are all divergent balanced positions of the compactly supported period basis through 16 factors; the factor-count cutoff is computational, and larger half-powers outside that basis slice are not claimed complete)
How they were obtained:

Chuang–Yu, Corollary 28 expresses each selected finite part as a signed sum of exponentially convergent ordinary moments. Complex Arb quadrature in SageMath 10.9 at 460 bits integrates that combined kernel after t=exp(-u) on 0<=u<=128, and directly on 1<=t<=128, using panels of length at most four and absolute/relative tolerance 2^(-400).

more

The omitted endpoints are bounded explicitly. From [3], I_0(t)<=exp(t), K_0(t)<=log(2/t)+1 for 0<t<=1, and K_0(t)<=sqrt(pi/(2t))*exp(-t) for t>0. The generator integrates the resulting positive elementary majorants exactly apart from enclosing ball arithmetic, adds their absolute bounds to the quadrature result, then propagates the period conversion. This encloses the defined finite part through the proved sum rule, without assigning a remainder bound to a divergent asymptotic series. The computed balls retain 394–403 bits of relative accuracy before checkpoint serialization; the stored decimals give 100 significant digits. Saved ball checkpoints conservatively include decimal rounding of both center and radius.

A separate check evaluates the defining balanced integral, subtracts its specified polynomial counterterms, and adds the integrated asymptotic tail, at (cutoff, retained order)=(256,120) and (384,160), using tanh–sinh quadrature at 150 and 190 working decimal digits, with 1800-bit Arb midpoint Bessel evaluations. Both estimates agree with every one of the 20 stored-convention values within its last-place interval. The checker does not use the generating convergent sum rule; it shares Arb Bessel evaluation but uses a different quadrature. Its quadrature and asymptotic tail remainder are heuristic, so this independent check is corroboration, not the basis for the proven label. Known-answer controls integral t*K_0(t)^2 dt=1/2 and integral t^3*K_0(t)^2 dt=1/3 pass; one-percent shifts and sign reversals are rejected for all entries. Executable code, source-fingerprinted checkpoints and detailed comparisons are attached. The default generator resumes those certified checkpoints; --recompute performs the full integrations. Its default enumeration and the completeness note describe the same 20 entries.