Timeline for Size of sets with complete double

Current License: CC BY-SA 4.0

32 events
when toggle format what by license comment
May 18, 2021 at 11:35 answer added Jukka Kohonen timeline score: 4
Jun 15, 2020 at 7:27 history edited CommunityBot
Commonmark migration
Nov 28, 2019 at 10:04 comment added Oliver Roche-Newton Here is a closely related question which may be of interest. Perhaps Lucia's argument below can give a non-trivial improvement there too? mathoverflow.net/questions/323294/…
Nov 1, 2019 at 13:13 comment added Hailong Dao Sure, I will do that. We probably would just refer to this question and answers in the example. Will probably take a few months before the paper is ready though.
Nov 1, 2019 at 12:54 comment added Lucia If you eventually write up your examples, please do link them here. I'd be curious to see how such problems come up in commutative algebra.
Nov 1, 2019 at 12:43 comment added Hailong Dao @Lucia: this arose when I tried to estimate certain algebraic invariants for a class of examples. So your answer (and everyone's contributions) are very helpful. Thanks. I am not sure if more could be done. It looks like a precise answer is not possible, and I don't really need very precise estimates, just to say: "for this class of examples, this invariant is hard to compute, but some rough estimates exist".
Nov 1, 2019 at 12:39 vote accept Hailong Dao
Nov 1, 2019 at 9:53 comment added Lucia Did this problem occur in some context in your work, or was it largely curiosity? My answer settles your Question 2 in the negative. Please let me know if you find anything there unclear, or if you were looking for something more.
Oct 29, 2019 at 21:13 history edited Lucia
edited tags
Oct 29, 2019 at 21:12 answer added Lucia timeline score: 11
Oct 21, 2019 at 6:36 comment added Seva "Perhaps one needs to really study the techniques there." - Absolutely.
Oct 21, 2019 at 0:47 comment added Hailong Dao @Seva: I looked at some literature on additive bases, and my question can be translated to basis for $2n$ but lies inside $[n]$. Unfortunately no paper that I came across seemed to address this particular problem, and all the current records use elements in the interval $[n+1, 2n]$, which does not help immediately. Perhaps one needs to really study the techniques there.
Oct 20, 2019 at 17:09 comment added Seva Here is yet another construction which gives $|S|=2(1+o(1))\sqrt{2n}$, at least for certain values of $n$. Suppose that $n=2^{m}$ with $m$ odd. Let $S=\{0\}\cup S_0\cup S_1$, where $S_0$ is the set of numbers in $[1,n]$ with all non-zero binary digits in the "even positions", and $S_1$ is the set of numbers in $[1,n]$ with all non-zero binary digits in the "odd positions". It is quite obvious that $S+S$ is the subset sum set of the set $\{0,1,2,4,\dotsc,2^m\}$, which is the interval $[0,2^{m+1}]=[0,2n]$. Also, for $m$ odd, we have $|S_0|=|S_1|=2^{(m+1)/2}$, whence $|S|=2\sqrt{2n}+1$.
Oct 20, 2019 at 15:30 comment added Hailong Dao @Seva: that looks very relevant, thanks.
Oct 20, 2019 at 13:58 comment added Seva The term to google is additive basis, or maybe try additive basis for interval.
Oct 20, 2019 at 13:07 answer added Aaron Meyerowitz timeline score: 3
Oct 20, 2019 at 6:07 comment added Hailong Dao What the heck is going on at $n=31$?
Oct 20, 2019 at 5:45 answer added RobPratt timeline score: 8
Oct 20, 2019 at 5:29 comment added Hailong Dao @RobPratt: thanks, that's great! Did you use any special algorithm/computer? I imagine a brute force search will get large very quickly.
Oct 20, 2019 at 5:19 comment added RobPratt Here are the values of $m(n)$ for $n \in\{1,\dots,50\}$: 2 ,3 ,4 ,4 ,5 ,5 ,6 ,6 ,7 ,7 ,8 ,8 ,8 ,9 ,9 ,9 ,10 ,10 ,10 ,10 ,11 ,11 ,12 ,12 ,12 ,12 ,12 ,13 ,13 ,13 ,14 ,13 ,14 ,14 ,14 ,14 ,15 ,15 ,15 ,15 ,16 ,16 ,16 ,16 ,16 ,16 ,17 ,17 ,17 ,17. It doesn't seem to be in the OEIS.
Oct 20, 2019 at 4:57 history edited Hailong Dao CC BY-SA 4.0
added 51 characters in body; edited title
Oct 20, 2019 at 4:49 comment added Gerry Myerson OK, so, that's taking $d$ roughly $\sqrt{n/2}$.
Oct 20, 2019 at 4:45 comment added Hailong Dao So it is about $2d+n/d$ which minimizes at $2\sqrt{2n}$!
Oct 20, 2019 at 4:44 comment added Gerry Myerson Does it? $\sqrt n$ up to $d$, $\sqrt n$ down to $n-d$, and $\sqrt n$ between $d$ and $n-d$, I thought.
Oct 20, 2019 at 4:41 comment added Hailong Dao I think that gets us to $2\sqrt{2n}$ actually.
Oct 20, 2019 at 3:12 comment added Gerry Myerson Never mind, I miscalculated again. How about everything from zero to $d$, multiples of $d$ up to $n-d$, and everything from $n-d$ to $n$? That's $3\sqrt n$, asymptotically.
Oct 19, 2019 at 22:00 comment added Hailong Dao @GerryMyerson: Can you say more on that modification?
Oct 19, 2019 at 21:55 comment added Gerry Myerson Yes, you're right. A modification gets $2\sqrt2\sqrt n$ asymptotically, which isn't good enough.
Oct 19, 2019 at 21:47 comment added Hailong Dao @GerryMyerson: I don't think so, if I understood your comment right. It would be hard to get things in the range $[2n-d...2n]$.
Oct 19, 2019 at 21:43 comment added Gerry Myerson Don't you get $2\sqrt n$ asymptotically by taking everything from zero to $d$, the integer part of $\sqrt n$, and then all multiples of $d$ up to $n$?
Oct 19, 2019 at 18:15 history edited YCor
edited tags
Oct 19, 2019 at 18:03 history asked Hailong Dao CC BY-SA 4.0