On analytic perturbations of a family of Feigenbaum-like equations (Q607293): Difference between revisions

That a solution exists for the Feigenbaum-Coullet-Tresser equation \[ \phi(x)= {1\over\lambda} \phi(\phi(\lambda x)) \] is well-known. Since that equation has been so important in the study of renormalization and universality for unimodular maps, it is natural to ask if solutions exist for perturbations of this equation. The author of this paper considers a family of Feigenbaum-like equations \[ \phi(x)= {1+ \varepsilon\over\lambda} \phi(\phi(\lambda x))-\varepsilon x+ \tau(x), \] where \(\varepsilon\) is a small real number and \(\tau(x)\) is a small analytic perturbation defined on some complex neighborhood of \((-1,1)\) and real on \(\mathbb{R}\). This family appears in the study of period-doubling renormalization for area-preserving maps (in the work of \textit{D. Gaidashev} and \textit{H. Koch}, Period-doubling in area-preserving maps: an associated one-dimensional problem preprint, available at \url{http://www.math.uu.se/~gaidash}]). The author proves the existence of solutions \((\phi,\lambda)\). A significant complication in the proof arises because of the \(-\varepsilon x+\tau(x)\) terms. To get around this obstacle, the author constructs a-priori bounds for unimodular functions that turn out to be quite tight. The author also obtains good bounds on the scaling parameter \(\lambda\). An additional benefit of the author's approach is a new proof of the existence of the Feigenbaum-Coullet-Tresser function.