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 #27 Apr 30 2023 14:15:20
%S 1,1,2,5,13,34,90,236,621,1629,4274,11193,29337,76818,201173,526730,
%T 1379178,3610804,9453695,24750281,64798235,169644626,444138288,
%U 1162770238,3044180080,7969770106,20865148382,54625676431,143011928942
%N a(1)=1; a(n+1) = Sum_{k=1..n} a(k) a(floor(n/k)).
%C 4 is the only composite number n such that a(n+1) = 3a(n) - a(n-1) and if n is a composite number greater than 4 then a(n+1) > 3a(n) - a(n-1). - _Farideh Firoozbakht_, Feb 05 2005
%H Vincenzo Librandi, <a href="/A097417/b097417.txt">Table of n, a(n) for n = 1..1000</a>
%F Ratio a(n+1)/a(n) seems to tend to 1 + Golden Ratio = 2.61803398... = 1 + A001622. - Mark Hudson (mrmarkhudson(AT)hotmail.com), Aug 23 2004
%F Satisfies the "partial linear recursion": a(prime(n)+1) = 3*a(prime(n)) - a(prime(n)-1). This explains why we get a(n+1)/a(n) -> 1 + phi. Also, lim_{n->oo} a(n)/(1 + phi)^n exists but should not have a simple closed form. - _Benoit Cloitre_, Aug 29 2004
%F Limit_{n->oo} a(n)/(1 + phi)^n = 0.108165624886204570982244311730754895284041534583990405146651275318889227986... - _Vaclav Kotesovec_, May 28 2021
%p a[1]:=1: for n from 1 to 50 do: a[n+1]:=sum(a[k]*a[floor(n/k)],k=1..n): od: seq(a[i],i=1..51) # Mark Hudson, Aug 21 2004
%t a[1] = 1; a[n_] := a[n] = Sum[ a[k]*a[Floor[(n - 1)/k]], {k, n - 1}]; Table[ a[n], {n, 29}] (* _Robert G. Wilson v_, Aug 21 2004 *)
%o (PARI) {m=29;a=vector(m);print1(a[1]=1,",");for(n=1,m-1,print1(a[n+1]=sum(k=1,n,a[k]*a[floor(n/k)]),","))} \\ _Klaus Brockhaus_, Aug 21 2004
%Y Cf. A038044, A078346, A097919.
%K easy,nonn
%O 1,3
%A _Leroy Quet_, Aug 19 2004
%E More terms from _Klaus Brockhaus_, _Robert G. Wilson v_ and Mark Hudson (mrmarkhudson(AT)hotmail.com), Aug 21 2004