Prev: wgUrlProtocols="http\\:\\/\\/|https\\:\\/\\/|ftp\\:\\/\\/|irc\\:\\/\\/|gopher\\:\\/\\/|telnet\\:\\/\\/|nntp\\:\\/\\/|worldwind\\:\\/\\/|mailto\\:|news\\:|svn\\:\\/\\/",
Next: P = NP if and only if there exists a Turing machine T and a natural number k such that <j>(ri) < nk,
From: In-Betweener on 7 Jul 2010 17:57 Em 07/07/2010 05:31, William Elliot wrote: > On Tue, 6 Jul 2010, In-Betweener wrote: > (...) > > Exercise. R is the union of the transitive interiors of R. > > ---- Thanks to all of you. Weeks ago, I found that such a transitive interior was not unique, but after choosing one of them for a particular relation, I got so impressed with the resemblance of concepts that I forgot totally that I was not dealing with a unique "interior". Shame on me. :-) By the way, I avoided the intersection definition of R+ because the hypothetical R- couldn't be defined this way, nor using union, of course. The lattice structure explained by William enriches a lot the range of insights I can have with my "transitive toys". Best regards.
From: In-Betweener on 31 Jul 2010 10:17 Em 06/07/2010 19:59, In-Betweener wrote: > (...) Hi all. I'm back after some weeks off-line. In the meanwhile, I worked a little on the "transitive interior" concept. My main interest in the concept was on how to keep the transitivity of a relation after removing an arbitrary pair of it. By the way, I've found a work akin to the subject at http://www.cs.uu.nl/research/techreps/repo/CS-1987/1987-25.pdf, but it concentrates on transitive closure and *reduction*, besides to be concerned mostly about efficiency. Here are some results I found myself. Suppose R is a transitive relation and S is R - {<a,b>}. It is provable that S has at most two "transitive interiors", defined as below: T1 = S \ {<a,y> | ySb} T2 = S \ {<x,b> | aSx} Additionally, it is simple to define a third, unique, transitive relation, intersecting the two above: T = T1 /\ T2 I chose to call T *the* "transitive core" of S. For an arbitrary relation R, the transitive core, possibly empty, is the intersection of all transitive interiors of R. If R is already transitive, the transitive closure, the (unique, in this case) transitive interior and the transitive core are all the same. I would appreciate any comments or suggestions about the matter and the proposed terminology. I've found nothing on the subject so far. Thanks in advance.
First
|
Prev
|
Pages: 1 2 Prev: wgUrlProtocols="http\\:\\/\\/|https\\:\\/\\/|ftp\\:\\/\\/|irc\\:\\/\\/|gopher\\:\\/\\/|telnet\\:\\/\\/|nntp\\:\\/\\/|worldwind\\:\\/\\/|mailto\\:|news\\:|svn\\:\\/\\/", Next: P = NP if and only if there exists a Turing machine T and a natural number k such that <j>(ri) < nk, |