login

Year-end appeal: Please make a donation to the OEIS Foundation to support ongoing development and maintenance of the OEIS. We are now in our 61st year, we have over 378,000 sequences, and we’ve reached 11,000 citations (which often say “discovered thanks to the OEIS”).

a(n) = A298825(n)/n.
5

%I #119 Mar 13 2023 06:13:41

%S 1,0,0,1,0,0,0,2,-1,0,0,0,0,0,0,3,0,0,0,0,0,0,0,0,-1,0,-2,0,0,0,0,4,0,

%T 0,0,-1,0,0,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,5,0,0,

%U 0,0,0,0,0,-2,0,0,0,0,0,0,0,0,-3,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,0,0,0,-2

%N a(n) = A298825(n)/n.

%C From _Mats Granvik_, Apr 06 2019: (Start)

%C Positions of nonzero entries appear to be given by A001694.

%C The limit lim_{k->oo} Sum_{n=1..k} a(n)/n appears to converge to some number.

%C Sum_{n=1..30000} a(n)/n = 1.31897... This appears to be in agreement with:

%C (1/30000)*Sum_{n<=30000} log(A014963(n))*log(A014963(n+2)) = 1.30351...

%C If the limit can be proven to converge to a number greater than 1, then it is true that Sum_{n<=X} log(A014963(n))*log(A014963(n+2)) > X, where ">" means "greater than" as in usual inequality notation.

%C The twin prime conjecture, according to Terence Tao on his blog, is that Sum_{n<=X} log(A014963(n))*log(A014963(n+2)) >> X where ">>" means "asymptotically greater than". There he also says that the first Hardy Littlewood conjecture states that Sum_{n<=X} log(A014963(n))*log(A014963(n+h)) = S(h)*X + o(X), where "S(h)" is the singular series.

%C Compare this to the prime number theorem which is lim_{X->oo} (Sum_{n<=X} log(A014963(n)))/X = 1.

%C (End)

%H Antti Karttunen, <a href="/A298826/b298826.txt">Table of n, a(n) for n = 1..2500</a>

%H Mats Granvik, <a href="https://mathoverflow.net/q/308990/25104">Arithmetic properties of a sum related to the first Hardy-Littlewood conjecture</a>.

%H Terence Tao, <a href="https://terrytao.wordpress.com/2017/07/06/correlations-of-the-von-mangoldt-and-higher-divisor-functions-i-long-shift-ranges/">Correlations of the von Mangoldt and higher divisor functions I. Long shift ranges</a>. See second formula.

%F a(n) = Sum_{k=1..n} (A298674(k, n))/n = A298825(n)/n.

%F Conjecture: a(n) = (-1)^(n+1) * Sum_{d|n} A076479(d). - _Daniel Suteu_, Apr 04 2019

%F From _Mats Granvik_, Apr 06 2019: (Start)

%F The Dirichlet generating function, after Daniel Suteu above and Álvar Ibeas in A076479, appears to be: Sum_{n>=1} a(n)/n^s = zeta(s)^2*(Product_{j>=1} (1 - 2*prime(j)^(-s)))*(1 + Sum_{n>=2} ((1/2)/2^(n*(s - 1)))).

%F Conjectured formula: Let b(n) = 4*A056911(n) and c(n) = A000005(n)*A008683(n) + Sum_{j=1..length(b(1..N))} [b(j)=n]*A008836(n)*2/3) then a(n) = Sum_{k=1..n}[k|n] c(n/k)*A000005(k). (End)

%F Conjecture: a(n) = (-1)^(n+1) * Sum_{d|n} mu(d)*tau(d)*tau(n/d). - _Ridouane Oudra_, Nov 19 2019

%F The conjectured Dirichlet generating function simplifies to: Sum_{n>=1} a(n)/n^s = zeta(s)^2*(Product_{j>=1} (1 - 2*prime(j)^(-s)))*(1 + 2^(1 - s)/(2^s - 2)). - _Steven Foster Clark_, Sep 12 2022

%F Conjecture: abs(a(n)) = A361430(n). - _Vaclav Kotesovec_, Mar 12 2023

%t nn = 90;

%t A = Table[Boole[Mod[n, k] == 0], {n, nn}, {k, nn}];

%t B = Table[If[Mod[k, n] == 0, MoebiusMu[n]*n, 0], {n, nn}, {k, nn}];

%t T = (A.B);

%t TwinMangoldt = Table[a = T[[All, kk]];

%t F1 = Table[If[Mod[n, k] == 0, a[[n/k]], 0], {n, nn}, {k, nn}];

%t b = T[[All, kk + 2]];

%t F2 = Table[ If[Mod[n, k] == 0, b[[n/k]], 0], {n, nn}, {k, nn}];

%t (F1.F2)[[All, 1]], {kk, nn - 2}];

%t TT = Transpose[TwinMangoldt];

%t Table[Sum[TT[[n, k]], {k, n}]/n, {n, nn - 2}]

%t (* From _Mats Granvik_, Mar 17 2019: (Start) *)

%t nn = 108; b = 4*Select[Range[1, nn, 2], SquareFreeQ]; bb = Table[DivisorSigma[0, n]*(MoebiusMu[n] + Sum[If[b[[j]] == n, LiouvilleLambda[n]*2/3, 0], {j, 1, Length[b]}]), {n, 1, nn}];

%t cc = Table[Sum[If[Mod[n, k] == 0, bb[[n/k]]*DivisorSigma[0, k], 0], {k, 1,

%t n}], {n, 1, nn}]

%t (* This faster conjectured program agrees with Antti Karttunen's precomputed \

%t list of numbers. *)(* (End) *)

%t (* Dirichlet generating function *) s=7; nn=2500; N[Zeta[s]^2*Product[(1 - 2 Prime[j]^(-s)), {j, 1, nn}]*(1 + Sum[1/2/2^(n*(s - 1)), {n, 2, nn}]), 40] (* _Mats Granvik_, Apr 06 2019 *)

%o (PARI)

%o up_to = 256;

%o DirConv(ma,h) = { my(u = matsize(ma)[1], md = matrix(u,u)); for(n=1,u-h,for(k=1,u,md[n,k] = sumdiv(k,d,ma[n,d]*ma[n+h,k/d]))); (md); };

%o A298826list(up_to) = { my(h=2, matA = matrix(up_to+h,up_to+h,n,k,!(n%k)), matB = matrix(up_to+h,up_to+h,n,k,(!(k%n))*moebius(n)*n), matT = matA*matB, matD = DirConv(matT,2)); vector(up_to,i,(1/i)*sum(j=1,i,matD[j,i])); };

%o v298826 = A298826list(up_to);

%o A298826(n) = v298826[n]; \\ _Antti Karttunen_, Sep 30 2018

%Y Cf. A298674, A298824, A298825, A001694, A361430.

%K sign

%O 1,8

%A _Mats Granvik_, Jan 27 2018

%E More terms from _Antti Karttunen_, Sep 30 2018