Sur la perfection de l'ensemble de Julia des fonctions entières transcendantes

Claudio Meneghini
2008-08-17
Abstract: We show, in an elementary way, that the Julia set of one-complex-variable entire functions is nonempty and perfect.
Complex Variables
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to prove that the Julia set of a transcendental entire function on a complex variable is neither empty nor contains isolated points, that is, it is perfect. Specifically: 1. **Non - emptiness**: Prove that for any transcendental entire function \(f\), its Julia set \(J_f\) is non - empty, that is, \(J_f\neq\emptyset\). 2. **Perfection**: Prove that \(J_f\) is a perfect set, that is, it has no isolated points. ### Main Results - **Theorem 3**: For any transcendental entire function \(f\), its Julia set \(J_f\) is non - empty. - **Proof Idea**: - Utilize Lemma 2 and assume that \(f\) has at least one fixed point. - Discuss in two cases: - If \(f\) has two or more fixed points, then there exist at least two Fatou components \(C_1\) and \(C_2\), so \(\partial C_i\subset J_f\), thus \(J_f\neq\emptyset\). - If \(f\) has only one fixed point \(p\), assume that \(F_f = \mathbb{C}\), then \(p\) is either an attracting fixed point or a neutral fixed point. In both cases, by the König linearization theorem, it can be deduced that \(f\) is injective in a certain region, which contradicts the fact that \(f\) is a transcendental function. - **Theorem 4**: For any transcendental entire function \(f\), its Julia set \(J_f\) is perfect. - **Proof Idea**: - First, prove that \(J_f\) is infinite. If \(J_f\) contains at least two points, then according to Picard's theorem, \(f^{- 1}(J_f)\) is infinite and \(f^{-1}(J_f)\subset J_f\). - Assume that \(J_f\) contains only one point \(x\), then \(f(x)=x\) and \(f^{-1}(x)=\{x\}\). Consider the function \(g(z)=\frac{f(z)-x}{z - x}\), which is a transcendental entire function that does not take the value 0. According to Picard's theorem, it takes the value 1 at infinitely many points, so \(f\) has infinitely many fixed points. If infinitely many of these fixed points are in \(J_f\), then the conclusion holds; otherwise, there are infinitely many Fatou components, which also means that \(J_f\) is infinite. - Finally, use Hurwitz's lemma to prove that any point \(\xi\) in \(J_f\) is not an isolated point. ### Related Lemmas - **Lemma 1** (Zalcman's Renormalization Lemma): Let \(H = \{f_\alpha\}\) be a family of meromorphic functions on a domain \(W\subset\mathbb{C}\). If \(H\) is not normal at \(v\in W\), then there exist sequences \(\{v_n\}\to v\), \(\{r_n\}\downarrow0\), \(\{f_n\}\subset H\) and a non - constant meromorphic function \(h\) (whose spherical derivative is bounded on \(\mathbb{C}\)), such that \(\{f_n(v_n + r_n z)\}\) converges uniformly on every compact set of \(\mathbb{C}\) to \(h\). - **Lemma 2**: For any transcendental entire function \(f\), there exists \(\eta\in\mathbb{C}\) such that \(f^2(\eta)=\eta\). - **Proof**: Assume that such a point does not exist, then