A note on the homology of signed posets (Q1915856)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A note on the homology of signed posets |
scientific article |
Statements
A note on the homology of signed posets (English)
0 references
1 July 1996
0 references
Let \((P, \leq)\) be a poset with \(P = \{1, 2, \dots, n\}\). For each \(i \in P\) let \(e_i \) denote the \(i\)th unit coordinate vector in \(R^n\). Let \(S\) be the set of all \(e_i - e_j\) such that \(i < j\). Then \(S\) is a subset of the root system \(A_n\) and we have: (a) \(S \cap (- S) = \emptyset\); (b) \(\overline S \cap A_n = S\) (where \(\overline S\) is the positive linear closure of \(S\), i.e., the span of all linear combinations of vectors in \(S\) with nonnegative real coefficients). The notion of signed poset was introduced by \textit{V. Reiner} [J. Comb. Theory, Ser. A 62, 324-360 (1993; Zbl 0773.06008)] as a \(B_n\)-analogue of the construction above. Namely, a signed poset is a subset \(S\) of the root system \(B_n\) such that conditions (a) and (b) hold with \(A_n\) replaced by \(B_n\). There is a natural partial ordering \(P(S)\) associated to \(S\), and the homology of \(S\) is to be the homology of a certain subcomplex of the chain complex of \(P(S)\). One may hope that if a signed poset is \(EL\)-labellable then its homology be zero except in the top dimension, just as it holds for posets. Unfortunately, this conjecture is false. Nevertheless, the author proves that if \(P(S)\) is an \(EL\)-labellable poset (in particular, if \(P(S)\) is Cohen-Macaulay) and \(S\) has rank \(n\), then the homology of \(S\) vanishes for degrees outside the interval \([n/2,n]\).
0 references
Cohen-Macaulay poset
0 references
\(EL\)-labellable poset
0 references
root system
0 references
signed poset
0 references
homology
0 references
chain complex
0 references