A new approach to numerical computation of Hausdorff dimension of iterated function systems: applications to complex continued fractions (Q1990900)

The authors declare that in this paper the main interest is ``describing methods which give rigorous estimates for the Hausdorff dimension of invariant sets for (possibly infinite) iterated function systems or IFS's''. The attention is given to the case of invariant sets for IFS's associated to complex continued fractions. The present article deals with a Banach space of \(C^m\) functions. However, the authors note that the particular problem in this paper can also be set up in a Banach space of analytic functions in two complex variables. Some general well-known facts about iterated function systems, the Hausdorff dimension are described. Techiques using in proofs are explained by the authors in detail. Surveys on related results are given. In addition, the problem on computing the Hausdorff dimension of certain sets is discussed. The concept of an infinitesimal similitude is defined. The authors remark that this concept generalizes the concept of affine linear similitudes, which are affine linear contraction maps \(\theta\) satisfying for all \(x, y \in\mathbb R^n\) \[ d(\theta(x), \theta(y)) = cd(x, y), c<1. \] The main items of this research are following: -- The computation of the Hausdorff dimension of some invariant sets arising from complex continued fractions is considered. -- Some results concerning existence of \(C^m\) positive eigenfunctions for a class of positive (in the sense of order-preserving) linear operators are described. The problem on how one can often obtain explicit bounds on partial derivatives of the positive eigenfunctions, is considered. -- Explicit bounds on the partial derivatives of eigenfunctions of operators in which the mappings \(\theta_b\) are given by Möbius transformations which map a given bounded open subset H of \(\mathbb C := \mathbb R^2 \) into H, are obtained. These results are used for obtaining results about the case of infinite IFS's which are adequate for immediate purposes of this paper. -- Some spectral properties of the approximating matrices which justify standard numerical algorithms for computing their spectral radii are checked. -- The log convexity of the spectral radius\( r(L_s)\), which is exploited in the present numerical approximation scheme are discussed. Finally, one can note the following statements from the abstract of the authors: `` In our context, \(L_s\) (a parametrized family of positive, linear, Perron-Frobenius operators) is studied in a space of \(C^m\) functions and is not compact. Nevertheless, it has a strictly positive Cm eigenfunction vs with positive eigenvalue \(\lambda_s\) equal to the spectral radius of \(L_s\). Under appropriate assumptions on the IFS, the Hausdorff dimension of the invariant set of the IFS is the value \(s = s_{*}\) for which \(\lambda_s = 1\). To compute the Hausdorff dimension of an invariant set for an IFS associated to complex continued fractions, (which may arise from an infinite iterated function system), we approximate the eigenvalue problem by a collocation method using continuous piecewise bilinear functions. Using the theory of positive linear operators and explicit a priori bounds on the partial derivatives of the strictly positive eigenfunction \(v_s\), we are able to give rigorous upper and lower bounds for the Hausdorff dimension \(s_{*}\), and these bounds converge to \(s_{*}\) as the mesh size approaches zero. We also demonstrate by numerical computations that improved estimates can be obtained by the use of higher order piecewise tensor product polynomial approximations, although the present theory does not guarantee that these are strict upper and lower bounds. An important feature of our approach is that it also applies to the much more general problem of computing approximations to the spectral radius of positive transfer operators, which arise in many other applications.''