Gauss-Manin systems, Brieskorn lattices and Frobenius structures. I. (Q1415562): Difference between revisions

The construction of a Frobenius structure on the germ of the base space of the miniversal deformation of an isolated singularity is done using the theory of primitive forms of \textit{K. Saito} [Publ. Res. Inst. Math. Sci. 19, 1231--1264 (1983; Zbl 0539.58003)] and Hodge theory of Brieskorn lattices due to \textit{M. Saito} [Ann. Inst. Fourier 39, No. 1, 27--72 (1989; Zbl 0644.32005)]. A complete exposition of such a construction is given by \textit{C. Hertling} in [``Frobenius manifolds and moduli spaces for singularities'', Cambridge Tracts in Mathematics. 151. Cambridge: Cambridge University Press (2002; Zbl 1023.14018)]. Recently there have been developments in direction of generalizing such a construction, which has a local nature, to certain functions on affine varieties. The articles under review generalize such a construction to the case of convenient and non-degenerate Laurent polynomials [see \textit{A. G. Kouchnirenko}'s work, Invent. Math. 32, 1--31 (1976; Zbl 0328.32007)] which include the restriction of the linear form \(f=w_0u_0+\cdots+w_nu_n\) to the torus \(\prod_{i}u_i^{w_i}=1\), where \(w_i\) is a sequence of positive integers with greatest common divisor equal to \(1\). The case \(w_1=\cdots=w_n=1\) is already treated by \textit{S. Barannikov} in [``Semi infinite Hodge structures and mirror symmetry for projective spaces'', preprint] and this is the main source of inspiration for the authors. The first article (this paper) is mainly dedicated to the theoretical aspects in the construction of the Frobenius structure attached to a regular tame function \(f\) on an affine manifold. The first section developes some notions in \({\mathcal D}\)-modules theory such as partial Fourier transform, its regularity, relation with partial microlocalization, etc. In this section the authors use some unpublished result of B. Abdel-Gadir and B. Malgrange, whose proofs are presented in Appendix A. In the second section, the notion of a tame polynomial, associated Gauss-Manin system and the Malgrange-Kashiwara filtration are introduced and some results on holomorphicity, regularity, duality, etc. are proved. In the third section the universal unfolding of \(f\) is defined by asking that the Kodaira-Spencer map is an isomorphism. This gives the \(F\)-manifold structure on the deformation space \(X\). Then the authors uses K. Saito's infinitesimal period mapping to transport the connection and the pairing constructed from M. Saito's solution to Birkhoff's problem (the Appendix B discusses this problem) and K. Siato's primitive form. The fourth section is dedicated to the construction of the Frobenius structure in the case of a Laurent polynomial \(f\). For an \(f\) which is non-degenerate with respect to its Newton polyhedron and is convenient, they define the Newton filtration on the corresponding Brieskorn lattice and prove that it coincides with the filtration induced by Malgrange-Kashiwara filtration. They conclude that the spectrum of the Newton filtration is contained in \([0,n]\) and is symmetric with respect to \(\frac{n}{2}\). The second article (part II of this paper [the authors, aspects of Mathematica E36, 1--18 (2004; Zbl 1079.32014), see the paper below]) emphasizes on the Laurent polynomial \(f\) obtained from the example cited above. It is convenient and non-degenerate with respect to its Newton polyhedron. They calculate explicitly the spectral polynomial attached to \(f\). Its roots are rational numbers \(\sigma_w(k)\), \(k=0,1,\ldots,n\) constructed from \(w_i\)'s. Then, they describe the Frobenius structure on the universal unfolding of \(f\) in terms of a data (which consists of two matrices, one of them diagonal with entries \(\sigma_w(k)\), a non degenerate bilinear form and a vector in the space in which the matrices act) introduced by \textit{B. Dubrovin} [Lect. Notes Math. 1620, 120--348 (1996; Zbl 0841.58065)]. The calculations for \(f\) are quite explicit and it is recommended to the reader to follow the theoretical aspects of the first article and related calculations in the second article simultaneously.