30 Apr
2015
30 Apr
'15
3:34 p.m.
This paper at least has some interesting experimental facts about it: < http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.52.868&rep=rep1&typ... >. ——Dan
On Apr 30, 2015, at 1:59 PM, Allan Wechsler <acwacw@gmail.com> wrote:
Set a_k(0) = k, and a_k(n+1) = sigma(a_k(n)), where sigma(x) is the sum of the divisors of x.
My impression is that a_k grows just barely super-exponentially, with local chaos, but very smoothly overall (regardless of k).
Does anyone have any intuitions about whether a_2(n) = a_5(m) for any n, m? That is, are the forward orbits of 2 and 5 completely disjoint?