Macaulay-lex rings (Q351409)

Let \(S = k[X_{1}, \dots , X_{n}]\) be the polynomial ring over a field \(k\), with the standard gradation \(S = \bigoplus_{d\geq 0}S_{d}\) and \(\deg(X_{i}) = 1\) for all \(i=1, \dots , n\). Let \(\mathrm{Mon}(\mathcal{A})\) denote the collection all monomials in \(\mathcal{A}\subseteq S\). Given \(u\in \mathrm{Mon}(S)\), the integer \(m(u)\) is the defined as \(m(u) = \max\{i \mid X_{i} \mid u\}\). A \textit{\(d\)-vector space} is a \(k\)-vector subspace of \(S_{d}\) and a \textit{\(d\)-monomial space} is a \(d\)-vector space spanned by monomials. A \(d\)-monomial space is said to have the \textit{lexsegment} property if whenever \(X^{a}\in V_{d}\), \(X^{a} <_{\mathrm{lex}} X^{b}\) and \(X^{b}\in S_{d}\), we have \(X^{b}\in V_{d}\). An ideal \(I\) is called a \textit{piecewise lexsegment} ideal if it has the property that if \(G(I)\) is the minimal generating set of \(I\) and if \(z >_{\mathrm{lex}} w\in G(I)\) with \(\deg(z) = \deg(w)\) and \(m(z) \leq m(w)\), then \(z\in I\). A \(d\)-monomial space is called \textit{strongly-stable} if whenever \(u\in \mathrm{Mon}(V_{d})\) with \(X_{j} \mid u\), we have \(\frac{X_{i}u}{X_{j}}\in V_{d}\) for all for \(i < j\). A \(d\)-monomial space is called \textit{stable} if \(\frac{X_{i}u}{X_{m(u)}}\in V_{d}\) for all \(u\in \mathrm{Mon}(V_{d})\) and for all \(i< m(u)\). Let \(M\) be a monomial ideal in \(S\). The ideal \(M\) and the quotient ring \(S/M\) are said to be \textit{Macaulay-lex} if every Hilbert function of a graded ideal in the quotient ring \(S/M\) is attained by a lex ideal in \(S/M\). The theorem proved by \textit{D. A. Shakin} [Sb. Math. 194, No. 11, 1701--1724 (2003); translation from Mat. Sb. 194, No. 11, 117--140 (2003; Zbl 1076.13008)] states that a strongly stable ideal \(B\) of \(S\) is Macaulay-lex if and only if \(B\) is a piecewise lexsegment ideal. The article under review has generalized Shakin's theorem by proving that a stable ideal \(I\) of \(S\) is Macaulay-lex if and only if \(I\) is a piecewise lexsegment ideal. \textit{F. S. Macaulay} [Proc. Lond. Math. Soc. (2) 26, 531--555 (1927; JFM 53.0104.01)] proved that for every graded ideal \(J\) in \(S\), there exists a lex ideal \(L\) in \(S\) with the same Hilbert function. The theorem of \textit{G. F. Clements} and \textit{B. Lindström} [J. Comb. Theory 7, 230--238 (1969; Zbl 0186.01704)] proves that if \(2 \leq e_{1} \leq \cdots \leq e_{n} \leq \infty\) be a sequence of integers or \(\infty\) and if \(I = \langle X_{1}^{e_{1}}, \dots , X_{n}^{e_{n}} \rangle\), where \(X_{i}^{\infty} = 0\), then, for every graded ideal \(J\) in \(S/I\) there exists a lex ideal \(L\) in \(S/I\) with the same Hilbert function. The article under review has given simple proofs of the Macaulay's theorem and the Clements-Lindström theorem and has generalized the Clements-Lindström theorem by proving the following result: Theorem. Let \(I = \langle X_{1}^{e_{1}}, X_{1}^{t_{1}}X_{2}^{e_{2}}, X_{1}^{t_{1}}X_{2}^{t_{2}}X_{3}^{e_{3}},\dots , X_{1}^{t_{1}}\cdots X_{n-1}^{t_{n-1}}X_{n}^{e_{n}}\rangle\), where \(2 \leq e_{1} \leq e_{2} \leq \cdots \leq \infty\), \(t_{i} < e_{i}\) for all \(i\) and \(X_{i}^{\infty} = 0\). \(I\) is Macaulay-lex if and only \(I\) has one of the following properties: {\parindent=6mm \begin{itemize}\item[(1)] \(I = \langle X_{1}^{e_{1}}, X_{1}^{t_{1}}X_{2}^{e_{2}}, \dots , X_{1}^{t_{1}}X_{n}^{e_{n}}\rangle\), \item[(2)] \(I = \langle X_{1}^{e_{1}}, X_{1}^{e_{1}-1}X_{2}^{e_{2}}, X_{1}^{e_{1}-1}X_{2}^{t_{2}}X_{3}^{e_{3}},\dots , X_{1}^{e_{1}-1}X_{2}^{t_{2}}\cdots X_{n-1}^{t_{n-1}}X_{n}^{e_{n}}\rangle\). \end{itemize}}