home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!howland.reston.ans.net!usc!news.cerf.net!nic.cerf.net!jcbhrb
- From: jcbhrb@nic.cerf.net (Jacob Hirbawi)
- Newsgroups: sci.math
- Subject: Re: Combinatorical problem
- Date: 13 Jan 1993 04:10:26 GMT
- Organization: CERFnet Dial n' CERF Customer Group
- Lines: 70
- Distribution: world
- Message-ID: <1j04niINNe98@news.cerf.net>
- NNTP-Posting-Host: nic.cerf.net
-
- In article <greil.726829446@guug.de> greil@guug.de (Anton Greil) writes:
-
- > I'd like to append with a more general question:
- > T(n) forms a generating system for S(n) (for all n).
- > How can the combinatorics of this generating system be characterized?
- > Especially: does there exist a normal form of each generated element from
- > S(n) over the base T(n) ?
-
- I know of a form that has a number of nice features. (I'm not sure how it
- relates to the two forms suggested by Vaughan Pratt but chances are good it's
- a close relative). I'll look at permutations of degree 4 but all the results
- are easily extended:
-
- First let the transpositions T(4)={a1,b1,b2,c1,c2,c3} be defined by:
-
- a1 = (1,2), b1 = (1,3), c1 = (1,4), ...
- b2 = (2,3), c2 = (2,4),
- c3 = (3,4),
- then every permutation can be written uniquely as
-
- (ai^ri)(bj^rj)(ck^rk) with ri,rj,rk = 0 or 1
- and i=1,j=1 or 2,k=1,2, or 3, ...
-
- Inspite of the clumsy notation I hope the pattern is clear enough.
- This is equivalent to expressing the numbers from 0 to n!-1 in the
- "factorial base" . I'll do the first 10 permutations, (according to an
- obvious order):
-
- 1234 <--> (000) <--> (identity)
- 2134 <--> (100) <--> a1
- 1324 <--> (010) <--> b1
- 3124 <--> (110) <--> a1 b1
- 2314 <--> (020) <--> b2
- 3214 <--> (120) <--> a1 b2
- 1243 <--> (001) <--> c1
- 2143 <--> (101) <--> a1 c1
-
- p <--> (C2,C3,C4)
-
- The i'th coefficient, Ci, for permutation p can be calculated directly as the
- Ci = number of j < i such that p(j) > p(i)
-
- The sum of the the Ci's is odd or even according to the signature of p = -1,1.
-
- > And - excuse please, yet more general - is there a connection to the general
- > Moebius groups of the Euclidean spaces R_n, where each such Moebius transfor-
- > mation can be generated by reflections, which are also elements of the
- > (group-) order 2 ?
-
- There is a connection which is either superficial or very deep (sometimes it's
- hard to distinguish between the two ;-) ). These are also connected to
- semi-simple Lie algebras. Anyway Vaughan's decomposition of the symmetric
- group into even and odd permutations extends very nicely to all reflection
- groups. The symmetric group, S(n), is in fact the "simplest" reflection group
- and S(n+1) is associated with the (Weyl) reflection group of the Lie algebra
- su(n). All reflection groups have a subgroup of index 2, called their
- rotation group. The cosets of this subgroup divide the group into two subsets
- satissfying exactly the conditions of the original problem posted by
- Matthew Newman:
-
- >> For which n can one find a set B(n) such that for all p in S(n)
- >> either p is in B(n), or p differs by exactly one transposition
- >> from an element of B(n), and such that no two elements from B(n)
- >> differ by exactly one transposition ?
-
- Most of H.S.M Coxeter's books is a good reference on this as well as:
- "Finite Reflection Groups" by L.C.Grove and C.T.Benson.
-
- Jacob Hirbawi
- JcbHrb@CERF.net
-