home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!munnari.oz.au!manuel.anu.edu.au!dubhe.anu.edu.au!nunki.anu.edu.au!not-for-mail
- From: jmr@cs.anu.edu.au (Mike Robson)
- Newsgroups: comp.theory
- Subject: Re: Silence on P=NP
- Date: 24 Nov 1992 10:40:25 +1100
- Organization: Computer Science Department, ANU, Australia
- Lines: 53
- Distribution: inet
- Message-ID: <1erq59INNonj@nunki.anu.edu.au>
- References: <1992Nov21.222755.18863@sophia.smith.edu> <1992Nov23.091136.7123@fwi.uva.nl> <1erksgINNom1@nunki.anu.edu.au>
- NNTP-Posting-Host: nunki.anu.edu.au
-
- jmr@cs.anu.edu.au (Mike Robson) writes:
-
- >leen@fwi.uva.nl (Leen Torenvliet) writes:
-
- >>orourke@sophia.smith.edu (Joseph O'Rourke) writes:
-
- >>>The silence on the Swart & Gismondi proof of P=NP is deafening.
- >>>Will no one venture an opinion? My knowledge of linear programming
- >>>is insufficient to the task.
-
- >>So is about everybody's it seems. I guess we're all waiting for some
-
- > If noone out there has understood the argument well enough yet to accept it
- > or point out a flaw, why don't we pool our efforts to try and elucidate it?
-
- > Here's my comment to start the ball rolling:
- > I am trying to understand the Graph Isomorphism paper first since that is the
- > basis for the other two; I am fine as far as the end of page 5 where they talk
- > about `a larger convex set $C_1$'. This seems to be the definition of $C_1$
- > which crops up again significantly but I don't know which `larger convex set'
- > it is supposed to be. Anybody like to clarify?
-
-
- Let me follow up my own article with another query:
-
-
- At the top of page 10 we are told (without proof) that every solution to
- F3 can be represented as a linear combination of the $q$ vectors (with sum
- of coefficients equal to 1). Should this be obvious to me or is it the
- conclusion of some deep argument?
-
- It seems to me that the space spanned by these $q$ vectors should have the
- same dimension as that spanned by the original permutation matrices, namely
- (n-1)^2+1 (since any n by n matrix is in the space provided all row and column
- sums are the same). But we know (page 8) that F3 is a formulation involving
- OMEGA(n^6) variables and O(n^5) constraints so I would expect its solution space
- to have dimension OMEGA(n^6) if it isn't empty.
-
- What is the simple point I am missing here?
-
-
-
- >Dr J. M. (Mike) Robson,
- >Department of Computer Science,
- >Australian National University,
- >GPO Box 4,
- >Canberra, ACT 2601,
- >Australia.
-
- >Tel: [61] 6-2494001
- >Fax: [61] 6-2490010
- >email: jmr@cs.anu.edu.au
-
-