Optimal shadows and ideals in submatrix orders (Q5937930): Difference between revisions

Let \(P\) be a ranked poset with associated partial order \(\leq\), and denote its \(i\)th level (the set of all elements of rank \(i\)) by \(N(i,P)= N(i)\). For \(x\in P\), let \(\Delta(x)= \{y\in N(i-1)\mid y\leq x\}\) denote the shadow of \(x\). If \(<\) denotes a total (linear) order on \(P\) and if for \(X\subseteq N(i)\), \(C(X)\) is the set of first \(|X|\) elements of \(N(i)\) \((\text{wrt}<)\), then \(X\subseteq N(i)\) is compressed if \(X= C(X)\). If \(\Delta(C(X))\subseteq C(\Delta(X))\) for all \(i\), \(X\subseteq N(i)\), then \((P,\leq,\prec)\) is a Kruskal-Katona type. If \((P,\leq)\) admits a Kruskal-Katona type, it is a Macaulay poset. Identifying Kruskal-Katona types, i.e., producing \(\prec\) when \(\leq\) is known, is of much combinatorial interest and referred to as the Shadow Minimization Problem. It is the origin of much poset-combinatorial technique, a considerable quantity due to this author, who is active in this area of poset theory. In this particular case, the class of posets is given by: \(A\), \(B\) disjoint non-empty sets, \(P(A,B)= \{F\subseteq A\cup B\mid F\cap A\neq\emptyset\neq F\cap B\}\) with rank function \(r(F)=|F|-2\) and \(\leq\) denoting set inclusion, which can also be realized as a submatrix arrays \((F\cap A)\times (F\cap B)\) of \(A\times X\), for which the design of \(\prec\) to obtain the Kruskal-Katona type is the crux of the paper, the proof that it is so, the greater body and a verification of Sali's conjecture with an application to finding particular ideals in orders such as these above an extra.