A solution to Schröder's equation in several variables

Abstract: Let phi be a self-map of B^n, the unit ball in C^n, fixing 0, and having full-rank at 0. If phi (0)= 0, Koenigs proved in 1884 that in the well- known case n = 1, Schroeder's equation, f circ phi = phi '(0) f has a solution f, which is bijective near 0 precisely when phi '(0) eq 0. In 2003, Cowen and MacCluer formulated the analogous problem in C^n (for a non-negative integer n) by defining Schroeder's equation in several variables as F circ phi = phi '(0)F and giving appropriate assumptions on phi . The 2003 Cowen and MacCluer paper also provides necessary and sufficient conditions for an analytic solution, F taking values in C^n and having full-rank near 0 under the additional assumption that phi '(0) is diagonalizable. The main result of this paper gives necessary and sufficient conditions for a Schroeder solution F which has full rank near 0 without the added assumption of diagonalizability. More generally, it is proven in this paper that the functional equation F circ phi = phi '(0)^k F with k a positive integer, is always solvable with an F whose component functions are linearly independent, but if k > 1 any such F cannot be injective near 0. In 2007 Enoch provides many theorems giving formal power series solutions to Schroeder's equation in several variables. It is also proved in this note that any formal power series solution indeed represents an analytic function on the ball.

Recommendations

Cites work

Cited in

(8)

This page was built for publication: A solution to Schröder's equation in several variables

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q265491)