11

I would like to know the current state of the phase transition for random k-sat, given n variables and m clauses, what is the best known c=m/n for upper and lower bounds.

Kaveh
  • 21,577
  • 8
  • 82
  • 183
Tayfun Pay
  • 2,598
  • 2
  • 18
  • 45
  • 2
    Did you try a reference search? cf. http://meta.cstheory.stackexchange.com/questions/300/how-to-ask-a-good-question – RJK Sep 30 '10 at 15:15
  • 3
    PS It looks to me like the second hit on Google is a freely-accessible article with answers to your question. (A 2009 arXiv article by Coja-Oghlan.) – RJK Sep 30 '10 at 15:23
  • See http://cstheory.stackexchange.com/questions/1130/what-does-one-mean-by-heuristic-statistical-physics-arguments/1229#1229 for a reasonably up to date perspective. Follow-ups by the people working in this area, such as Amin Coja-Oghlan and Dimitris Achlioptas, then have the answer you are looking for. – András Salamon Sep 30 '10 at 15:39
  • You may find this question useful: http://cstheory.stackexchange.com/questions/2168/how-many-instances-of-3-sat-are-satisfiable/2171#2171. In particular, see the first answer. – Giorgio Camerani Oct 25 '10 at 12:45
  • @Tayfun Pay: I am commenting here because you have deleted your questions and I can't comment there. Sorry, I didn't meant to be rude, I was asking for motivation to understand why you are interested in the question. In the other case, if you write down the definitions you will see the difference between them (many-one reductions does not need to preserve the number of certificates, it can map an instance of the problem A to an instance of problem B and there is no condition on the certificates at all). I think it would be better if you migrated the question to Math.SE. (more) – Kaveh Aug 03 '11 at 02:18
  • There is nothing wrong with moving a question to Math.SE, I myself sometimes ask questions that are not research level over there. Let me know if you decide to improve the deleted question or migrate the second one to Math.SE and I can undelete them for you (and give a more detailed answer for the second one over there). – Kaveh Aug 03 '11 at 02:21
  • See also subsequent question https://cstheory.stackexchange.com/questions/22016/has-there-been-any-research-on-k-sat-above-the-satisfiability-threshold for additional pointers to work exploring the phase transition. – András Salamon Feb 13 '15 at 09:53
  • see old franco survey. very different transition behavior on regular graph coloring cnfs. +daniel GREs2380 – daniel pehoushek Nov 09 '21 at 14:36

1 Answers1

20

Dimitris Achlioptas covers this in a survey article from the first edition of the Handbook of Satisfiability (PDF of draft).

There is conjectured to be a single threshold $r_k$ for each $k \ge 3$, so that when $m/n > r_k$ then a random $k$-SAT formula with $m$ clauses and $n$ variables is unsatisfiable with high probability, and so that when $m/n < r_k$ then a random $m$-clause, $n$-variable $k$-SAT formula is satisfiable with high probability. (More precisely, the conjecture is that in the limit as $n$ tends to infinity, the probability of satisfiability is 0 or 1 in these two regimes, respectively.)

Assuming that this Satisfiability Threshold Conjecture holds, the best known bounds for $r_k$ are

k                   3      4      5     7     10      20
Best upper bound 4.51  10.23  21.33 87.88 708.94 726,817
Best lower bound 3.52   7.91  18.79 84.82 704.94 726,809

(this table appears on the page indicated as 247 in the draft).

In a more recent manuscript (arXiv:1411.0650), Jian Ding, Allan Sly and Nike Sun showed that for all sufficiently large $k$, there is in fact a single threshold $r_k = 2^k\ln 2 - (1+\ln 2)/2 + o(1)$, and the error term $o(1)$ in this expression goes to zero as $k$ tends to infinity.

András Salamon
  • 19,000
  • 3
  • 64
  • 150