On local LYM identities (Q2441383): Difference between revisions

For a fixed natural number \(n\), let \([n]\) denote the set \(\{1, 2, \ldots, n\}\) and \(2^{[n]}\) denote the family of all subsets of \([n]\). If \(\emptyset\neq \mathcal{F}\subseteq 2^{[n]}\) and \(A\nsubseteq B\) for all \(A, B\in \mathcal{F}\), \(A\neq B\), then \(\mathcal{F}\) is called a Sperner family or an antichain. The well-know LYM inequality due to Lubell, Yamamoto and Meshalkin states that \[ \sum_{F\in \mathcal{F}}\frac{1}{{n\choose {\mid F\mid}}}\leq 1. \] Let \(\mathcal{P}_k\) be the set of all subsets of \([n]\) of cardinality \(k\). For each family \(\emptyset\neq \mathcal{A}\subseteq \mathcal{P}_k\), the upper shadow of \(\mathcal{A}\) is defined as \[ \nabla\mathcal{A}=\{A \subseteq [n] : | A|=k+1, A\supset B \;\text{for \;some} \;B\in \mathcal{A}\}. \] Kleitman and Harper proved that the LYM inequality is equivalent to the normalized matching property (or local LYM inequality), which says that for each \(\mathcal{F}\subseteq \mathcal{P}_k\) and every integer \(0\leq k < n\), \[ \frac{\mid\nabla\mathcal{F}\mid}{{n \choose {k+1}}}\geq \frac{\mid\mathcal{F}\mid}{{n\choose k}}. \] \textit{R. Ahlswede} and \textit{Z. Zhang} [Adv. Math. 80, No. 2, 137--151 (1990; Zbl 0705.05002)] lifted the LYM inequality to an identity, called the AZ identity. Thus, one expects that the same sharpening of the local LYM inequality is equivalent to the AZ identity. In the paper under review, the author introduces a local LYM identity which sharpens the local LYM inequality and proves that it is equivalent to the AZ identity.