Birational geometry of higher-dimensional Fano varieties (Q281765)

The aim of the survey under review is to give a complete account for the author's results on birational geometry of higher-dimensional Fano varieties \(V\) (cf. [\textit{A. V. Pukhlikov}, J. Reine Angew. Math. 688, 209--218 (1992; Zbl 0773.14004); Sb. Math. 193, No. 3, 445--471 (2002; Zbl 1024.14005); translation from Mat. Sb. 193, No. 3, 135--161 (2002); Izv. Math. 74, No. 5, 925--991 (2010; Zbl 1220.14014); translation from Izv. Ross. Akad. Nauk, Ser. Mat. 74, No. 5, 45--114 (2010)]). Chapter 1 is devoted to the method of maximal singularities. Here the author introduces the notions of the threshold of canonical adjunction \(c(V,\Sigma)\) of a movable linear system \(\Sigma\) on \(V\), its virtual analog \(c_{\text{virt}}(V,\Sigma)\), and introduces (recalls) the notion of birational (super) rigidity in the form \(c(V,\Sigma) = c_{\text{virt}}(V,\Sigma)\) for any \(\Sigma\). The author also considers an (effective) obstruction for this equality to occur. Namely, Propositions 1.1.2--1.1.6 setup generalities for the so-called \textit{maximal singularity} of the linear system \(\Sigma\), ending up eventually with the celebrated Noether-Fano inequality. Next the author goes to the problem of exclusion of maximal singularities. The main tool here is the \textit{\(4n^2\)-inequality} (stated in Theorem 1.2.1). The chapter concludes with applying the technique developed to a complete description of the group of birational automorphisms of a (generic) complete intersection of a quadric and a cubic \(V_{2 \cdot 3} \subset \mathbb{P}^5\) (cf. Theorem 1.3.1). In Chapter 2, the author describes the technique of hypertangent divisors, which is then applied to prove birational superrigidity of generic Fano complete intersection of index \(1\). In general, if \(X\) is an irreducible projective variety with some point \(o \in X\), an effective divisor \(D\) is called \textit{hypertangent} (with respect to \(o\)) if the \(H\)-degree of \(D\) is strictly less than the multiplicity of \(D\) at \(o\) (cf. Definition 2.1.1). Suppose now that \(V \subset \mathbb{P}^{M + k}\) is a general Fano complete intersection of dimension \(\geq 4\) and codimension \(\geq 2k + 1\). Then hypertangent divisors are used to establish the inequality \[ \frac{\text{mult}_oY}{\deg Y} \leq \frac{4}{\deg V} \] for any \(o \in X\) and irreducible subvariety \(Y \subset V\) (cf. 2.2.1). More precisely, one may assume that \(V\) is \textit{regular} at \(o\) (see Definition 2.2.1), in which case a sequence of hypertangent linear systems \(\Lambda_i\) on \(V\) is constructed inductively and the needed estimate follows (cf. Corollaries 2.2.1 and 2.2.2). Similar method is used to prove birational rigidity of Fano manifolds that are cyclic covers of other Fanos (see subsection \textbf{2.2.3} for details). The Chapter concludes with a proof that being a regular complete intersection is an open property (this is achieved via the codimension estimate of the respective ``irregular'' set). Chapter 3 studies birational superrigidity of iterated Fano double covers \(V\). These can be defined as complete intersections in the weighted projective space of type \(d_1\cdots d_k\cdot 2l_1\cdots 2l_m\) for some integers \(d_i,l_j\). Similarly as above, the author proves birational superrigidity of (generic) such \(V\) (cf. Theorems 3.1.2 and 3.1.3) under certain regularity assumptions introduced in Definition 3.1.2. In Chapter 4, the author addresses the question of birational (super)rigidity of certain singular hypersurfaces -- under some regularity condition as always (cf. Definition 4.1.2). The main result (Theorem 4.1.1) states basically that \(V\) is either birationally superrigid, or it is birationally rigid and the group \(\text{Bir}\,V = \mathbb{Z}\slash 2\mathbb{Z}\) is generated by an explicit (Galois) involution. The proof easily reduces to studying the case of singular point \(p \in V\) as (potential) maximal singularity (see Proposition 4.1.3). The latter is either not possible (when \(\text{mult}_p\,V \leq M - 3\)), or there exists a birational automorphism of \(V\) that ``untwists'' \(p\), so that the point is no longer a maximal singularity, yet such singularity appears on the blowup \(V_0 \longrightarrow V\) of \(p\) (see Proposition 4.1.4). The latter is achieved, again, by an elaborate construction of hypertangents and counting multiplicities (cf. 4.2). In the last Chapter 5, the author studies double covers \(V \longrightarrow \mathbb{P}^M, M \geq 5\), ramified at a general hypersurface \(W\) of degree \(2(M - 1)\). He proves that every rationally connected fiber space structure on \(V\) is induced by the linear projection onto \(\mathbb{P}^1\) from some projective space \(P \subset \mathbb{P}^M\) (see Theorem 5.1.1). This implies in particular that \(V\) is birationally superrigid. The strategy of the proof is to show that for any linear system \(\Sigma\), satisfying \(c(V,\Sigma) > c_{\text{virt}}(V,\Sigma)\), there exists a unique subspace \(P\) of codimension \(2\) such that \(\sigma^{-1}(V \cap P)\) is the maximal subvariety of \(V_P\), where \(\sigma: V_P \longrightarrow V\) is the blowup of \(V \cap P\) (cf. Proposition 5.1.2). Again the argument is based on elaborate study of maximal singularities (involving the additional \textit{\(8n^2\)-inequality} which is proved in subsection 5.4).