An analogue of Hilbert's theorem 90 for the ring of entire functions

From MaRDI portal
(Redirected from Publication:1122028)





The authors study the matrix equation \[ (*)\quad H(\lambda)H(\lambda L)H(\lambda L^ 2)...H(\lambda L^{m-1})=I, \] where \(L\in {\mathbb{C}}^{q,q}\) satisfies \(L^ m=I\) for some positive integer m and H: \({\mathbb{C}}^ q\to {\mathbb{C}}^{n,n}\) is entire in \(\lambda\). They show that (*) holds if and only if \(H(0)^ m=L\) and there exists \(E(\lambda)\in {\mathbb{C}}^{n,n}(\lambda)\), with both E and \(E^{-1}\) entire such that \(H(\lambda)=E(\lambda)H(0)E(\lambda L)^{-1}\). A condition comparable to the one in Hilbert's Theorem 90 is that \(H(\lambda)=W(\lambda)W(\lambda L)^{-1}\), where W is meromorphic, but the authors do not assume this. They approach the problem by introducing a new scalar parameter, deriving a linear system of differential equations for a related quantity, and proving a somewhat more general proposition from which their main result follows. They also generalize this result in several very interesting ways and give other interpretations for it. In the final part of the paper, they apply the main result to the study of the Stokes phenomena for second order linear differential equations with polynomial coefficients, thinking of the coefficients as parameters. They show that the central connection coefficients are entire functions of the parameters.












This page was built for publication: An analogue of Hilbert's theorem 90 for the ring of entire functions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1122028)