Studying the multiple Mellin--Barnes integrals by means of multidimensional residues (Q1972621): Difference between revisions

By a multiple Mellin-Barnes integral we mean the integral \[ \frac{1}{(2\pi i)^n}\int\limits_{\gamma+i\mathbb R^n} \frac{\prod\limits_{j=1}^{m}\Gamma \biggl(\sum\limits_{\nu =1}^{n}a_{j\nu}z_\nu +b_j\biggr)} {\prod\limits_{k=1}^{p}\Gamma \biggl(\sum\limits_{\nu =1}^{n}c_{k\nu}z_\nu +d_k\biggr)} t_1^{-z_1}\dots t_n^{-z_n} dz_1\cdots dz_n,\tag{1} \] where \(\Gamma\) is the Euler gamma-function and \(a_{j\nu}, b_j, c_{k\nu}\), and \(d_k\) are reals. The main result of the present article claims that, if \(n=2\), then, given coefficients \(a_{j\nu}\) and \(c_{k\nu}\), we can construct a~half-space \(\Pi_\Delta\) in the Euclidean space \(\mathbb C^2\) of the integration variable \(z=(z_1,z_2)\) such that the integral (1) is expressed in terms of the sum of the residues of the integrated differential form at some points of the intersection of the hyperplanes \[ \sum\limits_{\nu =1}^{n}a_{j\nu}z_\nu +b_j=-\nu , \quad \nu=0,1,2,\dots . \] As a consequence of this fact we infer that, as a rule, the Mellin-Barnes integrals converge to the Horn series. Our study of integral (1) is based on an abstract multidimensional Jordan lemma by \textit{M.~Passare} and the authors [Contributions to complex analysis and analytic geometry. Based on a colloquium dedicated to Pierre Dolbeault, Paris, France, June 23-26, 1992. Braunschweig: Vieweg, 233-241 (1994; Zbl 0832.32011)].