I guess it is well known that the transition from 2-SAT to 3-SAT is the gap
between P and NP.

The transition from 3-SAT to 4-SAT must be polynomial, because 4-SAT is
obviously within SAT and thus is NP-complete.  You only need poly time to
translate from within NP-complete.  So the answer to (3) is yes.

The gap lies between 2-SAT and 3-SAT.  My impression is that 2-SAT can be
easily solved with convex optimization but not for 3-SAT, where *convexity*
is lost.

Think, intuitively:  a quadratic equation defines a parabola, that has only
at most 1 extremum.  Whereas, a cubic equation can be like a sideways "S"
shape, that has 2 points where the derivative is 0, but they are not global
extrema.  Thus, the area under the cubic curve is not *convex*, and the
usual (fast) method of convex optimization cannot apply.

The loss of convexity when going from degree 2 to 3 is precisely where P
becomes NP.

(But this is just my crude impression, I may have gotten some details
wrong, which I am currently studying...)



-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657
Powered by Listbox: http://www.listbox.com

Reply via email to