Parabolic curves of diffeomorphisms asymptotic to formal invariant curves (Q1637092): Difference between revisions

A parabolic curve for a holomorphic germ $f$ tangent to the identity at the origin in several complex variables is a one-dimensional holomorphic curve (with the origin in its boundary) which is $f$-invariant and where the dynamics reproduces the parabolic one-dimensional dynamics of a petal in the classical Leau-Fatou flower theorem. \textit{J. Écalle} [Publ. Math. Orsay 85--05, 585 p. (1985; Zbl 0602.30029)] and, with different methods, \textit{M. Hakim} [Duke Math. J. 92, No. 2, 403--428 (1998; Zbl 0952.32012)] proved the existence of parabolic curves tangent to the so-called non-degenerate characteristic directions, and that if a parabolic curve is tangent at the origin to some direction then that direction must be characteristic. A germ tangent to the identity always have characteristic directions, but they might all be degenerate; so it remains open the problem of existence of parabolic curves tangent to a given degenerate characteristic direction. \par This paper deals with the closely related problem of existence of a parabolic curve asymptotic to a (necessarily invariant) formal curve. In other words, instead of prescribing only the first tangent (the characteristic direction) of a parabolic curve all higher-order tangencies are prescribed too. This is relevant because 2-dimensional holomorphic germs tangent to the identity always admit a formal invariant curve (necessarily tangent to a possibly degenerate characteristic direction). \par The main result of this paper says that, in dimension 2, given a formal invariant curve not contained in the set of fixed points of a germ tangent to the identity $f$ then there always exists a parabolic curve for $f$ or $f^{-1}$ which is asymptotic to the formal curve. The proof consists of two steps. First, a reduction of the pair (germ, invariant formal curve) to a normal form by means of change of coordinates and blow-ups; then when the pair is in normal form the authors prove the existence of a parabolic curve for the germ (or its inverse) asymptotic to the formal curve, adapting Hakim's arguments. \par Finally, in the last section of the paper the authors outline a way to generalise their approach to germs tangent to the identity in $\mathbb{C}^n$ with $n\ge 3$.