Effective finite generation for adjoint rings (Q486743)

One of the central problems in birational geometry is to prove the finite generation of graded rings of the form \(R(X, D) :=\bigoplus_{n=0}^\infty H^0(\mathcal{O}_X(nD))\), where \(D\) is a Cartier divisor on a projective variety \(X\). \newline In fact, if such ring is finitely generated, then one could try to study the geometry of \(X\) by reducing to the study of the variety \(Y :=\mathrm{Proj}(R(X, D))\) and the study of the fibers of the natural rational map \(X \dashrightarrow Y\) induced by the Iitaka fibration of \(D\). After the seminal work of Birkar--Cascini--Hacon--McKernan [\textit{C. Birkar} et al., J. Am. Math. Soc. 23, No. 2, 405--468 (2010; Zbl 1210.14019)], it is now a standard result that if \(D \sim_\mathbb{Q} a(K_X+\Delta), \; a \in \mathbb{Z}_{>0}\) then \(R(X, D)\) is indeed finally generated if the pair \((X, \Delta)\) has mild singularities and either \(\Delta\) or \(K_X+\Delta\) are big. In view of this result, one may wonder if it is possible to provide effective bounds on the degree of the generators with respect to the standard grading. In particular, the existence of such a bound immediately provides also an effective way of computing the stable base locus of \(D, \; Bs(D):= \bigcap_{m \geq 0} |mD|\). The paper under review is concerned with this question for projective varieties of dimension 2 and 3. The main theorem states that on a smooth projective threefold \(X\) if \(D = a(K_X+\Delta), \; a \in \mathbb{Z}_{>0}\) is a Cartier divisor then the ring \(R(X, D)\) is generated by elements of degree at most \(5q\), where \(q\) is a natural number that can be effectively computed in terms of the rank of the Néron-Severi group of \(X\), provided that the pair \((X, \Delta)\) satisfies the assumptions of the previous paragraph and moreover \(\Delta\) is numerically effective. \newline In their proof, the authors consider a minimal model \(f : X \dashrightarrow Y\) for \((X, \Delta)\) and show that the index of the singularities that appear on \(Y\) can be bounded in terms of the rank of the Néron-Severi group. This nice observation then reduces the proof to a generalization of Kollár's Effective Base Point Free Theorem [\textit{J. Kollár}, Math. Ann. 296, No. 4, 595--605 (1993; Zbl 0818.14002)] together with the analysis of the multiplicative structure of \(R(X, D)\). The latter is carried out by means of a Mumford regularity-type of argument. \newline In the case of surfaces the authors obtain a similar but more general statement.