OFFSET
1,1
COMMENTS
By definition a(n)>n. Conjecture: a(n) is always defined. Often the b sequences for two values of n merge and a(n) is the same for both values. So some numbers, such as 69, 631, 1167 and 689027, occur in the sequence more often than others.
Is sum(k=1,n,a(k))/(n^2*log(n)) bounded?
LINKS
Robert Israel, Table of n, a(n) for n = 1..1800
MAPLE
f:= proc(n) local b, k;
b[1]:= n;
for k from 1 do
b[k+1]:= b[k] - trunc(k/b[k]);
if b[k+1] = 0 then return k+1 fi;
od;
end proc:
map(f, [$1..100]); # Robert Israel, Nov 13 2024
MATHEMATICA
trunc[x_] := If[x>0, Floor[x], Ceiling[x]]; a[n_] := Module[{k, b}, For[k=0; b=n, b!=0, k++, b-=trunc[k/b]]; k]
CROSSREFS
KEYWORD
AUTHOR
Benoit Cloitre, Aug 25 2002
EXTENSIONS
Edited by Dean Hickerson, Aug 26 2002
STATUS
approved