cp's OEIS Frontend

This is a front-end for the Online Encyclopedia of Integer Sequences, made by Christian Perfect. The idea is to provide OEIS entries in non-ancient HTML, and then to think about how they're presented visually. The source code is on GitHub.

Showing 1-6 of 6 results.

A220002 Numerators of the coefficients of an asymptotic expansion in even powers of the Catalan numbers.

Original entry on oeis.org

1, 5, 21, 715, -162877, 19840275, -7176079695, 1829885835675, -5009184735027165, 2216222559226679575, -2463196751104762933637, 1679951011110471133453965, -5519118103058048675551057049, 5373485053345792589762994345215, -12239617587594386225052760043303511
Offset: 0

Views

Author

Peter Luschny, Dec 27 2012

Keywords

Comments

Let N = 4*n+3 and A = sum_{k>=0} a(k)/(A123854(k)*N^(2*k)) then
C(n) ~ 8*4^n*A/(N*sqrt(N*Pi)), C(n) = (4^n/sqrt(Pi))*(Gamma(n+1/2)/ Gamma(n+2)) the Catalan numbers A000108.
The asymptotic expansion of the Catalan numbers considered here is based on the Taylor expansion of square root of the sine cardinal. This asymptotic series involves only even powers of N, making it more efficient than the asymptotic series based on Stirling's approximation to the central binomial which involves all powers (see for example: D. E. Knuth, 7.2.1.6 formula (16)). The series is discussed by Kessler and Schiff but is included as a special case in the asymptotic expansion given by J. L. Fields for quotients Gamma(x+a)/Gamma(x+b) and discussed by Y. L. Luke (p. 34-35), apparently overlooked by Kessler and Schiff.

Examples

			With N = 4*n+3 the first few terms of A are A = 1 + 5/(4*N^2) + 21/(32*N^4) + 715/(128*N^6) - 162877/(2048*N^8) + 19840275/(8192*N^10). With this A C(n) = round(8*4^n*A/(N*sqrt(N*Pi))) for n = 0..39 (if computed with sufficient numerical precision).
		

References

  • Donald E. Knuth, The Art of Computer Programming, Volume 4, Fascicle 4: Generating All Trees—History of Combinatorial Generation, 2006.
  • Y. L. Luke, The Special Functions and their Approximations, Vol. 1. Academic Press, 1969.

Crossrefs

The logarithmic version is A220422. Appears in A193365 and A220466.
Cf. A220412.

Programs

  • Maple
    A220002 := proc(n) local s; s := n -> `if`(n > 0, s(iquo(n,2))+n, 0);
    (-1)^n*mul(4*i+2, i = 1..2*n)*2^s(iquo(n,2))*coeff(taylor(sqrt(sin(x)/x), x,2*n+2), x, 2*n) end: seq(A220002(n), n = 0..14);
    # Second program illustrating J. L. Fields expansion of gamma quotients.
    A220002 := proc(n) local recF, binSum, swing;
    binSum := n -> add(i,i=convert(n,base,2));
    swing := n -> n!/iquo(n, 2)!^2;
    recF := proc(n, x) option remember; `if`(n=0, 1, -2*x*add(binomial(n-1,2*k+1)*bernoulli(2*k+2)/(2*k+2)*recF(n-2*k-2,x),k=0..n/2-1)) end: recF(2*n,-1/4)*2^(3*n-binSum(n))*swing(4*n+1) end:
  • Mathematica
    max = 14; CoefficientList[ Series[ Sqrt[ Sinc[x]], {x, 0, 2*max+1}], x^2][[1 ;; max+1]]*Table[ (-1)^n*Product[ (2*k+1), {k, 1, 2*n}], {n, 0, max}] // Numerator (* Jean-François Alcover, Jun 26 2013 *)
  • Sage
    length = 15; T = taylor(sqrt(sin(x)/x),x,0,2*length+2)
    def A005187(n): return A005187(n//2) + n if n > 0 else 0
    def A220002(n):
        P = mul(4*i+2 for i in (1..2*n)) << A005187(n//2)
        return (-1)^n*P*T.coefficient(x, 2*n)
    [A220002(n) for n in range(length)]
    
  • Sage
    # Second program illustrating the connection with the Euler numbers.
    def A220002_list(n):
        S = lambda n: sum((4-euler_number(2*k))/(4*k*x^(2*k)) for k in (1..n))
        T = taylor(exp(S(2*n+1)),x,infinity,2*n-1).coefficients()
        return [t[0].numerator() for t in T][::-1]
    A220002_list(15)

Formula

Let [x^n]T(f(x)) denote the coefficient of x^n in the Taylor expansion of f(x) then r(n) = (-1)^n*prod_{i=1..2n}(2i+1)*[x^(2*n)]T(sqrt(sin(x)/x)) is the rational coefficient of the asymptotic expansion (in N=4*n+3) and a(n) = numerator(r(n)) = r(n)*2^(3*n-bs(n)), where bs(n) is the binary sum of n (A000120).
Also a(n) = numerator([x^(2*n)]T(exp(S))) where S = sum_{k>=1}((4-E(2*k))/ (4*k)*x^(2*k)) and E(n) the Euler numbers A122045.
Also a(n) = sf(4*n+1)*2^(3*n-bs(n))*F_{2*n}(-1/4) where sf(n) is the swinging factorial A056040, bs(n) the binary sum of n and F_{n}(x) J. L. Fields' generalized Bernoulli polynomials A220412.
In terms of sequences this means
r(n) = (-1)^n*A103639(n)*A008991(n)/A008992(n),
a(n) = (-1)^n*A220371(n)*A008991(n)/A008992(n).
Note that a(n) = r(n)*A123854(n) and A123854(n) = 2^A004134(n) = 8^n/2^A000120(n).
Formula from Johannes W. Meijer:
a(n) = d(n+1)*A098597(2*n+1)*(A008991(n)/A008992(n)) with d(1) = 1 and
d(n+1) = -4*(2*n+1)*A161151(n)*d(n),
d(n+1) = (-1)^n*2^(-1)*(2*(n+1))!*A060818(n)*A048896(n).

A220411 The denominators of J. L. Fields generalized Bernoulli polynomials.

Original entry on oeis.org

1, -6, 60, -504, 2160, -3168, 786240, -51840, 1762560, -82736640, 1437004800, -858470400, 101896704000, -1881169920, 8659353600, -855305256960, 14071151001600, -150493593600, 8252165711462400, -70431001804800, 24434139856896000, -1076294062964736000
Offset: 0

Views

Author

Peter Luschny, Dec 30 2012

Keywords

Comments

See A220412 for definitions and references.

Programs

  • Mathematica
    F[0, ] = 1; F[n, x_] := F[n, x] = -2x Sum[Binomial[n-1, 2k+1] BernoulliB[2k+2]/(2k+2) F[n-2k-2, x], {k, 0, n/2-1}] // Expand;
    a[n_] := (-1)^n Denominator[Together[F[2n, x]]];
    Table[a[n], {n, 0, 21}] (* Jean-François Alcover, Jun 27 2019 *)

A239792 Numerator of b_{2n}(1/4), where b_{n}(x) are Nörlund's generalized Bernoulli polynomials.

Original entry on oeis.org

1, -1, 3, -61, 1261, -4977, 999645, -16820653, 288427601, -1975649524361, 250373334235999, -741069328361243, 2017175162278526957, -16484758150014378103, 1866091048556360006871, -747145289541069391049541, 558035966935526487401599645, -94004035636878314426017611
Offset: 0

Views

Author

Peter Luschny, Mar 26 2014

Keywords

References

  • Y. L. Luke, The Special Functions and their Approximations, Vol. 1. Academic Press, 1969, page 34.
  • N. E. Nörlund, Vorlesungen über Differenzenrechnung, Berlin, 1924.

Crossrefs

Cf. A220412, A239793 (denominators).

Programs

  • Maple
    b := proc(n) option remember; if n < 1 then 1 else
    -add(binomial(n-1, k-1)*bernoulli(k)*b(n-k)/k, k= 2..n)/2 fi end:
    A239792 := n -> numer(b(2*n));
    seq(A239792(n), n=0..17);
  • Mathematica
    b[n_] := b[n] = If[n < 1, 1, -Sum[Binomial[n - 1, k - 1] BernoulliB[k] b[n - k]/k, {k, 2, n}]/2];
    a[n_] := b[2n] // Numerator;
    Table[a[n], {n, 0, 17}] (* Jean-François Alcover, Jun 28 2019, from Maple *)

Formula

Let b(n) = -Sum_{k=2..n} (C(n-1, k-1)*Bernoulli(k)*b(n-k)/k)/2 for n>0 and otherwise 1. Then a(n) = numerator(b(2*n)).

A239793 Denominator of b_{2n}(1/4), where b_{n}(x) are Nörlund's generalized Bernoulli polynomials.

Original entry on oeis.org

1, 24, 320, 10752, 184320, 360448, 23855104, 94371840, 285212672, 267764367360, 3720515420160, 987842478080, 201004469452800, 103903848824832, 637716744110080, 11997870882291712, 368450744514248704, 2251799813685248, 164633587978155851776, 9367487224930631680
Offset: 0

Views

Author

Peter Luschny, Mar 26 2014

Keywords

Comments

See A239792 for references.

Crossrefs

Cf. A220412, A239792 (numerators).

Programs

  • Maple
    b := proc(n) option remember; if n < 1 then 1 else
    -add(binomial(n-1, k-1)*bernoulli(k)*b(n-k)/k, k= 2..n)/2 fi end:
    A239793 := n -> denom(b(2*n));
    seq(A239793(n), n=0..19);
  • Mathematica
    b[n_] := b[n] = If[n < 1, 1, -Sum[Binomial[n - 1, k - 1] BernoulliB[k] b[n - k]/k, {k, 2, n}]/2];
    a[n_] := b[2 n] // Denominator;
    Table[a[n], {n, 0, 19}] (* Jean-François Alcover, Jun 28 2019 *)

Formula

Let b(n) = -sum_{2<=k<=n}(C(n-1, k-1)*Bernoulli(k)*b(n-k)/k)/2 for n>0 and otherwise 1. Then a(n) = denominator(b(2*n)).

A006934 A series for Pi.

Original entry on oeis.org

1, 1, 21, 671, 180323, 20898423, 7426362705, 1874409467055, 5099063967524835, 2246777786836681835, 2490122296790918386363, 1694873049836486741425113, 5559749161756484280905626951, 5406810236613380495234085140851, 12304442295910538475633061651918089
Offset: 0

Views

Author

Keywords

Comments

Formula (21) in Luke (see ref.): Let y = 4*n+1. Then for n -> oo
Pi ~ 4*(n!)^4*2^(4*n)/(y*((2*n)!)^2)*(sum_{k>=0}((-1)^k*y^(-2*k)* A006934(k)/A123854(k)))^2. (Luke does not reference the sequences in this form.) - Peter Luschny, Mar 23 2014
This might be related to the numerators of eq. (18) in N. Elezovic' "Asymptotic Expansions of Central Binomial...", J. Int. Seq. 17 (2014) # 14.2.1. - R. J. Mathar, Mar 23 2014
Several references give an erroneous value of 1874409465055 instead of a(7) in the formula for pi. - M. F. Hasler, Mar 23 2014

References

  • Y. L. Luke, The Special Functions and their Approximation, Vol. 1, Academic Press, NY, 1969, see p. 36.
  • N. J. A. Sloane and Simon Plouffe, The Encyclopedia of Integer Sequences, Academic Press, 1995 (includes this sequence).

Crossrefs

Programs

  • Maple
    A006934_list := proc(n) local k, f, bp;
    bp := proc(n,x) option remember; local k; if n = 0 then 1 else -x*add(binomial(n-1,2*k+1)*bernoulli(2*k+2)/(k+1)*bp(n-2*k-2,x), k=0..n/2-1) fi end:
    f := n -> 2^(3*n-add(i, i=convert(n,base,2)));
    add(bp(2*k,1/4)*binomial(4*k,2*k)*x^(2*k), k=0..n-1);
    seq((-1)^k*f(k)*coeff(%,x,2*k), k=0..n-1) end:
    A006934_list(15);  # Peter Luschny, Mar 23 2014
    # Second solution, without using Nörlund's generalized Bernoulli polynomials, based on Euler numbers:
    A006934_list := proc(n) local a,c,j;
    c := n -> 4^n/2^add(i, i=convert(n,base,2));
    a := [seq((-4)^j*euler(2*j)/(4*j), j=1..n)];
    expand(exp(add(a[j]*x^(-j), j=1..n))); taylor(%, x=infinity, n+2);
    subs(x=1/x, convert(%,polynom)): seq(c(iquo(j,2))*coeff(%,x,j), j=0..n) end:
    A006934_list(14); # Peter Luschny, Apr 08 2014
  • Mathematica
    A006934List[n_] := Module[{c, a, s, sx}, c[k_] := 4^k/2^Total[ IntegerDigits[k, 2]]; a = Table[(-4)^j EulerE[2j]/(4j), {j, 1, n}]; s[x_] = Series[Exp[Sum[a[[j]] x^(-j), {j, 1, n}]], {x, Infinity, n+2}] // Normal; sx = s[1/x]; Table[c[Quotient[j, 2]] Coefficient[sx, x, j], {j, 0, n}]];
    A006934List[14] (* Jean-François Alcover, Jun 02 2019, from second Maple program *)
  • Sage
    @CachedFunction
    def p(n):
        if n < 2: return 1
        return -add(binomial(n-1,k-1)*bernoulli(k)*p(n-k)/k for k in range(2,n+1,2))/2
    def A006934(n): return (-1)^n*p(2*n)*binomial(4*n,2*n)*2^(3*n-sum(n.digits(2)))
    [A006934(n) for n in (0..14)]  # Peter Luschny, Mar 24 2014

Formula

Let p(n,x) = sum(k=0..n, x^k*A220412(n,k))/A220411(n) then a(n) = (-1)^n*p(n,1/4)*A123854(n)*A001448(n). - Peter Luschny, Mar 23 2014
Pi = lim_{n->oo} 2^{4n+2}/((4n+1)*C(2n,n)^2)*(sum_{k=0..oo} (-1)^k*a(k)/(A123854(k)*(4n+1)^{2k}))^2. - M. F. Hasler, Mar 23 2014

Extensions

a(7) corrected, a(8)-a(14) from Peter Luschny, Mar 23 2014

A239795 a(n) = A239793(n)/2^(3*n).

Original entry on oeis.org

1, 3, 5, 21, 45, 11, 91, 45, 17, 1995, 3465, 115, 2925, 189, 145, 341, 1309, 1, 9139, 65, 2255, 148995, 108675, 1645, 270725, 21879, 583, 4389, 4959, 59, 1548729, 27027, 60775, 130985, 15525, 1065, 66047553, 2567565, 39, 2133, 56457, 1411, 8161615, 2639
Offset: 0

Views

Author

Peter Luschny, Mar 26 2014

Keywords

Comments

See A239792 for references.

Crossrefs

Programs

  • Maple
    b := proc(n) option remember; if n < 1 then 1 else
    -add(binomial(n-1, k-1)*bernoulli(k)*b(n-k)/k, k= 2..n)/2 fi end:
    A239795 := n -> denom(b(2*n))/2^(3*n):
    seq(A239795(n), n=0..43);

Formula

Let b(n) = -Sum_{2<=k<=n} (C(n-1, k-1)*Bernoulli(k)*b(n-k)/k)/2
for n>0 and otherwise 1. Then a(n) = denominator(b(2*n))/2^(3*n).
Showing 1-6 of 6 results.