SearcharxivSearch

arXiv subjects

Mutsumi Saito

Publications and source records attributed to Mutsumi Saito.

10 recordsLinked to original sources

Logarithmic $A$-hypergeometric series ${\textrm I}\! {\textrm I}\! {\textrm I}$

We study the logarithmic coefficients that can occur at a fixed fake exponent in an $A$-hypergeometric series subject to prescribed negative support conditions. Let $L=\operatorname{Ker}_{\mathbb Z}(A)$. For a fixed generic weight $w$, a fixed fake exponent $v_0$, and an ordered negative support family, we first derive a finite system of constant-coefficient differential equations whose solutions encode the admissible logarithmic coefficients. A normalization of the coefficient equations shows that, for $u\in L$, the normalized coefficient associated with $x^{v_0+u}$ depends only on the negative support of $v_0+u$. From this finite system, we identify the annihilator of the coefficient space with an explicitly defined colon ideal. For the negative support family determined by the direction $w$, we further identify this colon ideal with the primary component of the indicial ideal supported at $v_0$, shifted to the origin. We next introduce an ambient perturbation construction in which the fake exponent is perturbed in the full ambient space rather than only within the affine space $v_0+L_{\mathbb C}$. We prove that the ambient perturbation construction produces $A$-hypergeometric series and realizes the full coefficient space. Finally, we compare the ambient perturbation construction with the intrinsic perturbation construction developed in our previous papers. The intrinsic construction always yields a subspace of the full coefficient space, and it realizes the full coefficient space if and only if a natural equality between the corresponding colon ideals holds. We also give several sufficient conditions for this equality.

math.AG

Logarithmic A-hypergeometric series II

In this paper, following [6], we continue to develop the perturbing method of constructing logarithmic series solutions to a regular A-hypergeometric system. Fixing a fake exponent of an A-hypergeometric system, we consider some spaces of linear partial differential operators with constant coefficients. Comparing these spaces, we construct a fundamental system of series solutions with the given exponent by the perturbing method. In addition, we give a sufficient condition for a given fake exponent to be an exponent. As important examples of the main results, we give fundamental systems of series solutions to Aomoto-Gel'fand systems and to Lauricella's FC systems with special parameter vectors, respectively.

math.AG

Logarithmic A-hypergeometric series

The method of Frobenius is a standard technique to construct series solutions of an ordinary linear differential equation around a regular singular point. In the classical case, when the roots of the indicial polynomial are separated by an integer, logarithmic solutions can be constructed by means of perturbation of a root. The method for a regular A-hypergeometric system is a theme of the book by Saito, Sturmfels, and Takayama. Whereas they perturbed a parameter vector to obtain logarithmic A-hypergeometric series solutions, we adopt a different perturbation in this paper.

math.AG

Projective Linear Monoids and Hinges

Let V be a complex vector space. We propose a compactification PM(V) of the projective linear group PGL(V), which can act on the projective space P(V). After proving some properties of PM(V), we consider its relation to Neretin's compactification Hinge*(V).

math.RT

Limits of Jordan Lie subalgebras

Let g be a simple Lie algebra of rank n over C. We show that the n-dimensional abelian ideals of a Borel subalgebra of g are limits of Jordan Lie subalgebras. Combining this with a classical result by Kostant, we show that the g-module spanned by all n-dimensional abelian Lie subalgebras of g is actually spanned by the Jordan Lie subalgebras.

math.RT

The Freeness and Minimal Free Resolutions of Modules of Differential Operators of a Generic Hyperplane Arrangement

Let A be a generic hyperplane arrangement composed of r hyperplanes in an n-dimensional vector space, and S the polynomial ring in n variables. We consider the S-submodule D(m)(A) of the nth Weyl algebra of homogeneous differential operators of order m preserving the defining ideal of A. We prove that if n \geq 3, r > n,m > r - n + 1, then D(m)(A) is free (Holm's conjecture). Combining this with some results by Holm, we see that D(m)(A) is free unless n \geq 3, r > n,m < r - n + 1. In the remaining case, we construct a minimal free resolution of D(m)(A) by generalizing Yuzvinsky's construction for m = 1. In addition, we construct a minimal free resolution of the transpose of the m-jet module, which generalizes a result by Rose and Terao for m = 1.

math.CO

Noetherian properties of rings of differential operators of affine semigroup algebras

We consider the Noetherian properties of the ring of differential operators of an affine semigroup algebra. First we show that it is always right Noetherian. Next we give a condition, based on the data of the difference between the semigroup and its scored closure, for the ring of differential operators being anti-isomorphic to another ring of differential operators. Using this, we prove that the ring of differential operators is left Noetherian if the condition is satisfied. Moreover we give some other conditions for the ring of differential operators being left Noetherian. Finally we conjecture necessary and sufficient conditions for the ring of differential operators being left Noetherian.

math.RA

Primitive ideals of the ring of differential operators on an affine toric variety

Let $A$ be a $d\times n$ integer matrix whose column vectors generate the lattice $\Z^d$, and let $D(R_A)$ be the ring of differential operators on the affine toric variety defined by $A$. We show that the classification of $A$-hypergeometric systems and that of $\Z^d$-graded simple $D(R_A)$-modules (up to shift) are the same. We then show that the set of $\Z^d$-homogeneous primitive ideals of $D(R_A)$ is finite. Furthermore, we give conditions for the algebra $D(R_A)$ being simple.

math.RA

Logarithm-free A-hypergeometric series

We give a dimension formula for the space of logarithm-free series solutions to an A-hypergeometric (or a GKZ hypergeometric) system. In the case where the convex hull spanned by A is a simplex, we give a rank formula for the system, characterize the exceptional set, and prove the equivalence of the Cohen-Macaulayness of the toric variety defined by A with the emptiness of the exceptional set. Furthermore we classify A-hypergeometric systems as analytic D-modules.

math.AG

Isomorphism classes of A-hypergeometric systems

For a finite set A of integral vectors, Gel'fand, Kapranov and Zelevinskii defined a system of differential equations with a parameter vector as a D-module, which system is called an A-hypergeometric (or a GKZ hypergeometric) system. Classifying the parameters according to the D-isomorphism classes of their corresponding A-hypergeometric systems is one of the most fundamental problems in the theory. In this paper we give a combinatorial answer for the problem under the assumption that the finite set A lies in a hyperplane off the origin, and illustrate it in two particularly simple cases: the normal case and the monomial curve case.

math.AG