HARAOKA Yoshishige Kumamoto University, Department of Mathematics, Associate Professor, 理学部, 助教授 (30208665)
岡 幸正 熊本大学, 理学部, 助教授 (50089140)
KOHNO Mitsuhiko Kumamoto University, Department of Mathematics, Professor, 理学部, 教授 (30027370)
YAMAKI Hiroyoshi Kumamoto University, Department of Mathematics, Professor, 理学部, 教授 (60028199)
OKAMOTO Kazuo University of Tokyo, Graduate school of Mathematical Sciences, Professor, 大学院・数理科学研究科, 教授 (40011720)
YAMADA Kotaro Kumamoto University, Department of Mathematics, Associate Professor, 理学部, 助教授 (10221657)
|Budget Amount *help
¥3,700,000 (Direct Cost : ¥3,700,000)
Fiscal Year 1998 : ¥1,400,000 (Direct Cost : ¥1,400,000)
Fiscal Year 1997 : ¥2,300,000 (Direct Cost : ¥2,300,000)
The objecitve of this project is to study the general hypergeometric functions (GHF) which were introduced by us to give a unified understanding of the classical special functions such as Gauss hypergeometric, Kummer's confluent hypergeometric, Bessel, Hermite and Airy function and to give a natural generalization to the case of several variables.
1 : GHFs are defined as solutions of certain holonomic systems on the Grassmannian Gr_<r, n> and they have the integral representations in a formal sense whose integrand is a multivalued function on P^r. To obtain explicit resutis on GHF, it is important to understand this integral representation in the framework of de Rham theory, namely, as the dual pairing of cocycles and cycles of certain cohomology and homology groups. Here, for the integral on P^r, we defined the homology group as a locally finite homology group and then show that it is isomorphic to the relative homology group with compact supports for some pair of subsets P^r. Moreover
, using this result, we computed explicitly, in the case r=1, the dimension of the homology group and gave a basis of the group.
2 : For the Beta function B(alpha, beta), the simplest case of GHF with regular singularity, and for the Gamma function GAMMA(alpha), the simplest case of GHF with irregular singularity, the following formulas are well known :
B(alpha, beta)B(-alpha, -beta)=2pii(<@D71(/)alpha@>D7+<@D71(/)beta@>D7)(<@D7-e<@D12pii(alpha+beta)@>D1-1(/)e<@D12piialpha@>D1-1(e<@D12piibeta@>D1-1)@>D7), gamma(alpha)gamma(1-a)=<@D7pi(/)sinpialpha@>D7
We investigate the problem of understanding the above formulas from the viewpoint of de Rham theory. Explicitly we try to understand the right hand sides of the above formulas as a product of cohomological intersection number and the homological intersection number. For the GHF defined by the 1-dimensional integral, we computed explicitly the intersection matrix for the cohomoloy group by choosing its good basis.
By the choice of good basis, we can show that the intersection matrix turns out to be independent of the variables of the general hypergeometric function. The main reason for the computability of the intersection numbers is that the good basis has, at each singular point of the connection form of the de Rham complex, the analogous properties to the flat basis of the Jacobi ring for the simple singlarity of A-type. Less