2
$\begingroup$

Consider the set $\{1,2,\ldots,N\}$. Let $LCM(a,a')$ denote the lowest common multiple of the integers $a,a'.$

We say that $A\subset \{1,2,\ldots,N\}$ is $M-$good if $LCM(a,a')\geq M,$ for all $a\neq a' \in A.$

How large can the size $|A|$ be for an $M-$good set, as a function of $M,N.$ Specifically let $N\rightarrow \infty$ and $M=M(N)$ also go to infinity.

Is it possible to have, say, $M=N+1$ and $|A|\geq c N$, for some constant $c\in(0,1)$? If not, what is the order of $|A|$?

Also, how slowly must $M$ grow with $N$ to have $|A| \geq c N$, with $c \in (0,1)$?

$\endgroup$
1
  • $\begingroup$ A starting observation: If $A=\{(1-c)N, \dots, N\}$, then no two elements of $A$ have GCD more than $c N$, so we can take $M \geq N \frac{(1-c)^2}{c}$. This is probably improvable $\endgroup$ Commented Nov 6, 2015 at 0:16

2 Answers 2

4
$\begingroup$

You're asking for the size of a maximum independent set in the graph with vertices $\{1,2,\ldots,N\}$ and edges $(i,j)$ whenever $\text{lcm}(i,j) < M$.

I tried it in the case $M=N+1$ for $N$ up to $65$. It appears that the maximum $|A| = \lceil N/2 \rceil$, which is attained when $A = \{x: \lceil (N+1)/2 \rceil \le x \le N \}$.

$\endgroup$
2
  • $\begingroup$ Thanks. Did you use Mathematica or another package for the independent set? $\endgroup$
    – kodlu
    Commented Nov 6, 2015 at 4:03
  • $\begingroup$ I used the MaximumIndependentSet command in Maple's GraphTheory package. $\endgroup$ Commented Nov 6, 2015 at 4:50
5
$\begingroup$

Robert Israel's answer actually answers your first question with $c \geq 1/2$. I would only like to add that his example is actually the truth, in other words when $M=N+1$ then $c=1/2$.

Let us define a partially ordered set (poset) where $a \leq b$ if and only if $a$ divides $b$. Your $N+1$ good set must be an antichain in this poset. We can decompose this set into disjoint chains $\{x,2x,4x,8x, \ldots \}$ where $x$ is an odd number. The set $ \{ x : \lceil (N+1)/2 \rceil\ \leq x \leq N \}$ is an antichain with size equal to the number of odd numbers less than or equal to $N$. Since every antichain can have at most one element from a chain, Robert Israel's set is of optimal size.

Note that Dilworth's theorem (https://en.wikipedia.org/wiki/Dilworth%27s_theorem) guarantees that this strategy will always work when one would like to find an antichain of maximal size in a poset.

$\endgroup$
1
  • 1
    $\begingroup$ This is very nice. I'd like to accept both answers, but this is not possible, unfortunately. $\endgroup$
    – kodlu
    Commented Nov 6, 2015 at 4:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .