home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!sun4nl!and!jos
- From: jos@and.nl (Jos Horsmeier)
- Newsgroups: comp.theory
- Subject: Re: boolean satisfiability problem
- Message-ID: <3277@dozo.and.nl>
- Date: 22 Aug 92 11:10:31 GMT
- References: <Bt5Mz6.3xs@research.canon.oz.au>
- Organization: AND Software BV Rotterdam
- Lines: 39
-
- In article <Bt5Mz6.3xs@research.canon.oz.au> naylor@research.canon.oz.au (William Naylor) writes:
- |I am looking for information on algorithms to solve the
- |"boolean satisfiability problem". What is the largest problem that
- |can be solved routinely? What is the best known algorithm?
- |Any good references?
- |
- |The "boolean satisfiability problem" is a well-known NP-complete problem.
- |
- |A boolean function is expressed as a product of sums, for example
- |
- |F(a,b,c,d,...) = (a | ~b | c)&(~a | d | ~e)&...
- |
- |where the variables can take on the values TRUE and FALSE, and '|' represents
- |logical or, and '&' represents logical and. The problem to be solved is:
- |does any assignment of values to the variables cause the boolean function
- |F(...) to be TRUE?
-
- That reminds me ... Have a look at:
-
- Wang, Hao. `Towards Mechanical Mathematics'
- IBM J. Research Develop. VOl. 4
- No. 1. January 1960.
-
- This article describes exactly what you need: a method of deciding
- whether or not a WFF in the propositional calculus is a theorem.
-
- If you can find it (I think it's out of print for more than a decade
- now) you could have a look at:
-
- John McCarthy
- Lisp 1.5 Programmer's Manual
- The MIT Press
- ISBN 0 262 13011 4
-
- It describes a complete Lisp implementation of Wang's algorithm.
-
- kind regards,
-
- Jos aka jos@and.nl
-