Numbers of generators of perfect ideals (Q6085507)

Consider a local ring \((R, \mathfrak{m}, \mathbf{k})\) with residue field \(\mathbf{k} = R/\mathfrak{m}\). Let the Krull dimension of \(R\) be denoted by \(d\), and let \(I\) be an \(\mathfrak{m}\)-primary ideal of \(R\). The multiplicity of \(I\), denoted by \(e(I)\), is defined as: \[ e(I) = d! \lim_{n \to \infty} \frac{\ell(R/I^n)}{n^d}, \] where \(\ell\) represents the length function. The multiplicity of \(R\), denoted by \(e(R)\), is defined as \(e(\mathfrak{m})\). In his work [Adv. Math. 322, 940--970 (2017; Zbl 1391.13048)], \textit{L. Ma} stated what has been referred to as Conjecture 0.1 in the current paper under review: ``Let \((R, \mathfrak{m})\) be a complete local ring, and let \(J\) be a perfect ideal of \(R\). Then \(e(R/J) \geq e(R)\)''. The author claims that this conjecture seems difficult to attack, and addresses the situation when \(R\) is a regular local ring, seeking to formulate a theory specific to such rings. The number of generators for an ideal \(I\) in a regular ring is denoted by \(\mu(I)\). The author presents two additional questions concerning homogeneous ideals in graded rings: Question 0.2: Given a local ring \((R, \mathfrak{m})\) and a perfect ideal \(J\) of grade \(g > 0\), is it possible to conclude \[ \mu\left((J + \mathfrak{m}^n)/\mathfrak{m}^n\right) \leq \binom{g+n-2}{g-1}? \] Question 0.3: For a regular local ring \((R, \mathfrak{m})\) with a perfect ideal \(J\) of grade \(g > 1\), if \[ \mu\left((J + \mathfrak{m}^n)/\mathfrak{m}^n\right) \geq \binom{g+n-3}{g-2}, \] then can we conclude that \[ \mu(J) \leq \binom{g+n-2}{g-1}? \] Furthermore, the author provides the following definition: Let \((R, \mathfrak{m}, \mathbf{k})\) be a complete regular local ring. Choose a regular system of parameters \(x_{1}, \ldots, x_{d}\) for \(R\). Define an additive map \(\Phi_{n}: \mathfrak{m}^{n-1} \rightarrow \mathbf{k}[x_{1}, \ldots, x_{d}]\) whose image consists of all homogeneous elements of degree \(n-1\). For \(f \in \mathfrak{m}^{n-1}\), expressed as \(f = \sum c(t_{\alpha}, f) t_{\alpha}\) in Hironaka's representation, set \[ \Phi_{n}(f) = \sum_{\deg t_{\alpha} = n-1} \overline{c(t_{\alpha}, f)} t_{\alpha}, \] where \(\overline{c(t_{\alpha}, f)} \in \mathbf{k}\). The paper is replete with compelling results. Its main theorem states: Let \((R, \mathfrak{m}, \mathbf{k} )\) be a complete regular local ring of dimension \(d\), with an infinite residue field \(\mathbf{k}\), and suppose \(J\) is a perfect ideal of grade \(g \geq 2\). If \(x_{1}, \ldots, x_{d}\) constitute a regular system of parameters for \(R\), and assuming \(n > 2\) and that \(\Phi_{n}(J \cap \mathfrak{m}^{n-1})\) generates an ideal in \(\mathbf{k}[x_{1}, \ldots, x_{d}]\) of height at least \(g-1\), then \(\mu(J) \leq (n-1)^{g-1} + g-1\).