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”).
%I #33 Sep 27 2023 13:19:49
%S 2,3,3,0,0,9
%N Decimal expansion of Sum_{k>=2} (1/Sum_{j=2..k} j'), where n' is the arithmetic derivative of n.
%C Slow convergence.
%C a(7) is likely either 3 or 4. Is there a simple proof that this sum converges? - _Nathaniel Johnston_, May 24 2011
%C From _Husnain Raza_, Aug 29 2023: (Start)
%C The series indeed converges: we have that the series is C = Sum_{k>=2} (1/Sum_{j=2..k} A003415(j)).
%C Let s_k = Sum_{j=2..k} A003415(j) be the inner sum.
%C It is known that s_k = (1/2)*T_0*k^2 + O(k^(1+n)) for all real n > 0 where T_0 = A136141.
%C Therefore, 1/s_k = (2/T_0)*k^(-2) + O(k^(-3+n)) = (2/T_0)*k^(-2) + O(k^(-3)).
%C Summing both sides from k=2 to infinity, we have that:
%C C = Sum_{k >= 2} 1/s_k = Sum_{k >= 2} ((2/T_0)*k^(-2) + O(k^(-3))), which converges. (End)
%e 1/2' + 1/(2'+3') + 1/(2'+3'+4') + 1/(2'+3'+4'+5') + ... = 1 + 1/2 + 1/6 + 1/7 + ... = 2.33009...
%p with(numtheory);
%p P:=proc(i)
%p local a,b,f,n,p,pfs;
%p a:=0; b:=0;
%p for n from 2 to i do
%p pfs:=ifactors(n)[2];
%p f:=n*add(op(2,p)/op(1,p),p=pfs);
%p b:=b+f; a:=a+1/b;
%p od;
%p print(evalf(a,300));
%p end:
%p P(1000);
%t digits = 5; d[0] = d[1] = 0; d[n_] := d[n] = n*Total[Apply[#2/#1 &, FactorInteger[n], {1}]]; p[m_] := p[m] = Sum[1/Sum[d[j], {j, 2, k}], {k, 2, m}] // RealDigits[#, 10, digits]& // First; p[digits]; p[m = 2*digits]; While[Print["p(", m, ") = ", p[m]]; p[m] != p[m/2], m = 2*m]; p[m] (* _Jean-François Alcover_, Feb 21 2014 *)
%Y Cf. A003415, A190144, A190145, A190147.
%K nonn,more,cons
%O 1,1
%A _Paolo P. Lava_, May 05 2011
%E a(6) corrected and a(7) removed by _Nathaniel Johnston_, May 24 2011