Minimal free resolutions over complete intersections (Q896546)

The goal of this research monograph is to introduce the notion of a matrix factorization associated to a sequence of elements \(f_1, \dots, f_c\) of a local ring \(S\) and to use this notion to describe the eventual behavior of minimal free resolutions of modules over complete intersections. We provide a bit of historical background. Let \(f \in S\). A \textit{matrix factorization} of \(f\) is a pair of free \(S\)-modules \(F_0, F_1\) and a pair of \(S\)-linear maps \(d: F_0 \to F_1\) and \(h: F_1 \to F_0\) such that the compositions \(dh\) and \(hd\) are given by multiplication by \(f\). The notion of a matrix factorization was introduced by \textit{D. Eisenbud} in [Trans. Am. Math. Soc. 260, 35--64 (1980; Zbl 0444.13006)] as a tool to study the minimal free resolutions of modules over the ring \(S/(f)\). When \(S\) is regular and \(f\) is a regular element of \(S\), minimal free resolutions of modules over the hypersurface ring \(S/(f)\) are eventually 2-periodic, and their two-periodic tails are determined by matrix factorizations. Given a matrix factorization \((d, h)\) of \(f\), the cokernel of \(d\) is an \(S/(f)\)-module with a 2-periodic free \(S/(f)\)-resolution, and Eisenbud proves that every sufficiently high syzygy module over \(S/(f)\) may be obtained from a matrix factorization in this way. In Chapter 1 of the work under review, the authors give a definition of what they call a \textit{higher matrix factorization} of \(f_1, \dots, f_c \in S\), aiming to generalize the aforementioned work of Eisenbud. Set \(I\) to be the ideal \((f_1, \dots, f_c) \subseteq S\). Just as in the \(c = 1\) case, one may associate to a higher matrix factorization an \(S/I\)-module, which the authors call an \textit{HMF module}. In Chapter 2, matrix factorizations of one element are reviewed, and their relationship to minimal free resolutions over hypersurface rings, which we briefly recalled above, is described. Assume from now on that \(f_1, \dots, f_c\) is a regular sequence in \(S\). In Chapter 3, the authors construct a finite \(S\)-resolution of an HMF module associated to a higher matrix factorization of \(f_1, \dots, f_c\). In Chapter 4, the definition and some properties of \textit{CI-operators}, also known as \textit{Eisenbud operators}, are recalled. In Chapter 5, infinite free resolutions of HMF modules over the quotient \(S/I\) are constructed. When the HMF module arises from a \textit{minimal} higher matrix factorization (a notion defined in Chapter 1), the free resolutions constructed are minimal. In Chapter 6, it is proven that, when \(S\) is regular with infinite residue field and \(N\) is a finitely generated \(S/I\)-module, then every sufficiently high syzygy of \(N\) over \(S/I\) is an HMF module arising from a minimal higher matrix factorization of a generic choice \(f_1', \dots, f_c'\) of elements minimally generating \(I\) (see Corollary 6.4.3 of the work under review for a more precise formulation of this statement). The authors cite this result as the key motivation for their definition of a higher matrix factorization. \vskip\baselineskip In Chapter 7, the special case where \(S\) is Gorenstein is discussed. Finally, in Chapter 8, a definition of a \textit{morphism} of higher matrix factorizations is given. A morphism of higher matrix factorizations induces a morphism of HMF modules, and the authors prove that, if \(M\) and \(M'\) are \textit{stable syzygies} over \(S/I\) (as defined in Chapter 6) and \(S\) is Gorenstein, every morphism \(M \to M'\) arises from a morphism of higher matrix factorizations of \(f_1, \dots, f_c\).