The number $s_n$ of sum-free subsets of $[n]=\{ 1, 2, \dots n\}$ grows like
$c 2^{n/2}$. Subsets of odds in $[n]$ and large numbers in $[n]$ both
give rise to about $2^{n/2}$ sets: then there is a theorem due to Ben Green, giving an upper bound of $c 2^{n/2}$ so this is the true growth rate.
How big is $c$? Experimentally!
Is there a parity effect (even versus odd)?
Are there constants $c_e,c_o$ so that when $n$ is respectively odd versus even, $s_n/2^{n/2} \rightarrow c_0,c_e$?
|
|