Finite generation of canonical ring by analytic method (Q943399): Difference between revisions

In 2006 and 2007 the author posted two papers [A general non-vanishing theorem and an analytic proof of the finite generation of the canonical ring, \url{arXiv:math/0610740v1}; Additional explanatory notes on the analytic proof of the finite generation of the canonical ring, \url{arXiv:0704.1940}] containing the details of the techniques for an analytic proof of the finite generation of the canonical ring: If \(X\) is an \(n\)-dimensional complex projective manifold with canonical line bundle \(K_X\) and of general type in the sense that there exist \(m_0\in\mathbb N\) and \(c\in\mathbb R_+\) such that \(\dim_{\mathbb C}\Gamma(X,mK_X)\geq cm^n\) for \(m\geq m_0\), then the canonical ring \(\bigoplus_{m=1}^\infty\Gamma(X,mK_X)\) is finitely generated. In the present paper, the author gives an outline of a proof by refined analytic methods. They are based on those which he had used in the proof of the deformational invariance of the plurigenera of a smooth family of projective manifolds [Invent. Math. 134, No.~3, 661--673 (1998; Zbl 0955.32017)] and [\textit{I. Bauer} et al., Complex geometry. Collection of papers dedicated to Hans Grauert on the occasion of his 70th birthday. Berlin: Springer (2002; Zbl 0989.00069)], namely, the global generation of multiplier ideal sheaves, the extension theorem of \textit{T. Ohsawa} and \textit{K. Takegoshi} for L\(^2\) holomorphic functions [Math. Z. 195, 197--204 (1987; Zbl 0625.32011)] and a technique for getting L\(^2\) estimates for the \(\overline{\partial}\) operator. The main focus of the present paper lies on the description of the role of the so-called discrepancy subspace. For fixed bases \(s_1^{(m)},\dots,s_{q_m}^{(m)}\) of \(\Gamma(X,mK_X)\), \(m\in\mathbb N\), the author considers a function \(\varphi=\sum_{m=1}^\infty \varepsilon_m\sum_{j=1}^{q_m}| s_j^{(m)}| ^{\frac{m}{2}}\) and its partial sums \(\varphi_N=\sum_{m=1}^N \varepsilon_m\sum_{j=1}^{q_m}| s_j^{(m)}| ^{\frac{m}{2}}\), where \(\varepsilon_m\) are sufficiently small positive numbers to ensure convergence. Then \(\frac{1}{\varphi}\) and \(\frac{1}{\varphi_N}\) are metrics on \(K_X\), and \(P\in X\) is by definition a point of stable vanishing order if the quotient \(\frac{\varphi}{\varphi_N}\) is bounded in a neighborhood of \(P\) for some \(N\in\mathbb N\). The author applies a theorem of \textit{H. Skoda} on ideal generation [Ann. Sci. Éc. Norm. Supér. (4) 5, 545--579 (1972; Zbl 0254.32017)] in terms of multiplier ideals to prove that the canonical ring is finitely generated if every point of \(X\) is of stable vanishing order. The discrepancy subspace \(Z\) is the zero-set of a specific ideal sheaf \(\mathfrak I\) such that the pair \((Z,\mathfrak I)\) not only describes the set of points where the stable vanishing order is not achieved, but also measures the extend of the failure of reaching stable vanishing order in terms of the vanishing order of \(\mathfrak I\), simultaneously for all \(mK_X\) with \(m\) sufficiently large. This allows an induction process for the proof that \(Z=\emptyset\). Finally, the author discusses the relevance of these methods for a proof of the abundance conjecture. The Appendix of ten pages contains an exposition about multiplier ideal sheaves and closed positive currents on projective algebraic manifolds.