Title:Cooperation principle, stability and bifurcation in random complex dynamics

Abstract:We investigate random complex dynamics of rational or polynomial maps on the Riemann sphere. We show that regarding random complex dynamics of polynomials, generically, the chaos of the averaged system disappears at any point in the Riemann sphere due to the automatic coopeartion of many kinds of maps in the system, even though each map of the system has a chaotic part. Moreover, we show that (1) the set of stable systems is open and dense in the space of random dynamical systems of polynomials, and (2) for a stable system, there exist at most finitely many minimal sets and each minimal set is attracting. These (1) and (2) solve a random analogy of the most famous open conjecture in complex dynamics "hyperbolic rational maps are dense in the space of rational maps." Furthermore, we show that for a stable system, the orbit of a Hölder continuous function under the transition operator tends exponentially fast to the finite-dimensional space $U$ of finite linear combinations of unitary eigenvectors of the transition operator. From this, we obtain that the spectrum of the transition operator acting on the space of Hölder continuous functions has a gap between the set of unitary eigenvalues and the rest. Combining this with the perturbation theory for linear operators, we obtain that for a stable system constructed by a finite family of rational maps, the projection to the space $U$ depends real-analytically on the probability parameters. By taking a partial derivative of the function of probability of tending to a minimal set with respect to a probability parameter, we introduce a complex analogue of the Takagi function, which is a new concept. Many new phenomena which can hold in random complex dynamics but cannot hold in usual iteration dynamics of a single rational map are found and systematically investigated.