Random conformal snowflakes (Q990184)

The authors report on their search for extremal fractals. They argue that one should study random fractals instead of deterministic ones, introduce a new class of random fractals, called random conformal snowflakes, and investigate their properties. As a consequence, the authors significantly improve known estimates for the multifractal spectra of the harmonic measure. It is shown that, for this class, the average integral means spectrum, \(\overline\beta(t)\), is related to the main eigenvalue of a particular integral operator. Also, fractal approximation for this class is proved. An example of a snowflake is given. For this snowflake, it is proved that \(\overline\beta(t)> 0.2308\). This significantly improves the previously known estimate for the universal spectrum, \(B(1) > 0.17\), due to \textit{Ch. Pommerenke} [Univalent functions. With a chapter on quadratic differentials by Gerd Jensen. Studia Mathematica/Mathematische Lehrbücher. Band XXV. Göttingen: Vandenhoeck \& Ruprecht (1975; Zbl 0298.30014)].