home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cis.ohio-state.edu!zaphod.mps.ohio-state.edu!darwin.sura.net!wupost!gumby!yale!hsdndev!husc-news.harvard.edu!zariski!kubo
- From: kubo@zariski.harvard.edu (Tal Kubo)
- Newsgroups: sci.math
- Subject: Re: Counting Families of Subsets
- Keywords: Dedekind, unsolved
- Message-ID: <1992Nov15.220836.17488@husc3.harvard.edu>
- Date: 16 Nov 92 03:08:35 GMT
- References: <1992Nov15.152951.44802@kuhub.cc.ukans.edu>
- Organization: Dept. of Math, Harvard Univ.
- Lines: 23
- Nntp-Posting-Host: zariski.harvard.edu
-
- In article <1992Nov15.152951.44802@kuhub.cc.ukans.edu>
- kinnersley@kuhub.cc.ukans.edu (Bill Kinnersley) writes:
- >Take a base set S with N elements. There are 2^N subsets of S.
- >Let F be a family such that all elements of F are distinct subsets,
- >and no element of F is a subset of another.
- >
- >How many families like this are there, for given N?
-
- You are asking for D(n), the number of antichains of subsets of
- an N-element set.
-
- 1) Computing D(n) is an unsolved problem that goes back to
- Dedekind, if not earlier.
-
- 2) It has been proved that for large N, almost all the antichains
- have all their elements in the middle four layers (i.e. consist
- of sets of size n/2 - 2 to n/2+2).
-
- There is a journal called "Order", devoted to similar problems
- involving partially ordered sets. Check there for more recent
- information.
-
- -tk kubo@math.harvard.edu
-