The Noether inequality for Gorenstein minimal 3-folds (Q487959)

Given a smooth projective surface \(S\), it is possible to define the following numerical invariants: - \({K_S}^2\), the self intersection of \(K_S\); - \(p_g(S):= h^0(K_S) = h^2(\mathcal{O}_S)\), the geometric genus; - \(q(S) := h^1(\mathcal{O}_S)\), the irregularity; - \(\chi(S):= 1 - q(S) + p_g(S)\), the holomorphic Euler characteristic of \(S\). Surfaces of general type do not admit a nice classification as not general type ones, but [\textit{D. Gieseker}, Ann. Math. (2) 106, 45--60 (1977; Zbl 0381.14003)] proved that for every pair of integers \((a, b)\) there exists a coarse moduli space \(\mathcal{M}_{a, b}\) parametrizing minimal surfaces of general type with \(K^2 = a\) and \(\chi = b\). Therefore, it is natural to ask for which values of \(a\) and \(b\) is \(\mathcal{M}_{a, b}\) not empty. There are some restrictions given by the following inequalities: - Noether inequality: \(K^2 \geq 2\chi - 6 \geq 2 p_g - 4\); - Bogomolov-Miyaoka-Yau inequality: \(K^2 \leq 9\chi\); In 1980, Reid was the first that, underlining the importance of Noether inequality in the study of surface geography, asked the following question: What is the 3-dimensional version of the Noether inequality? The natural candidate \[ K_X^{\dim X} \geq 2(p_g - \dim X) \] was proved to be wrong in dimension three by \textit{M. Kobayashi} [J. Math. Soc. Japan 44, No. 1, 145--156 (1992; Zbl 0766.14033)]. But \textit{F. Catanese} et al. [Math. Res. Lett. 13, No. 4, 653--666 (2006; Zbl 1134.14024)] proved that for smooth minimal threefolds of general type the following inequality holds \(K_X \geq \frac{4}{3}p_g - \frac{10}{3}\). In the paper under review, the authors proved that the same inequality holds for a Gorenstein threefold \(X\) of general type. The difficult case that they have to consider is when \(X\) is canonically fibred by curves of genus 2 and the canonical linear system \(|K_X|\) has Gorenstein terminal singularities in the base locus. To remove this obstacle, they consider a special resolution of Gorenstein terminal singularities, called \textit{feasible Gorenstein resolution}.