Dynamical multifractal zeta-functions, multifractal pressure and fine multifractal spectra (Q2358386)

The author introduces some notions of multifractal topological pressure and dynamical multifractal zeta function. Roughly speaking this means to consider the points which are relevant for a given multifractal decomposition of the phase space. The dynamical setting is a conformal iterated function scheme and its representation by means of a symbolic space \((\Sigma,\sigma)\) with a finite alphabet. Thus if \(\varphi:\Sigma\to\mathbb R\) is a continuous map, \(X\) is a metric space and \(U:\mathcal M(\Sigma)\to X\) is a continuous functional then the upper and lower multifractal pressures are defined as \[ \begin{aligned}\overline{P}^U_C(\varphi) =\limsup_{n\to\infty}\frac{1}{n}\log\left(\sum\limits_{\overset{| i|=n}{UL_n[i]\subset C}}\sup_{u\in[i]}\exp\left(\sum\limits^{n-1}_{i=0}\varphi(\sigma^i(u))\right)\right) ,\\ \underline{P}^U_C(\varphi) =\liminf_{n\to\infty}\frac{1}{n}\log\left(\sum\limits_{\overset{| i|=n}{UL_n[i]\subset C}}\sup_{u\in[i]}\exp\left(\sum\limits^{n-1}_{i=0}\varphi(\sigma^i(u))\right)\right).\end{aligned} \] Here \([i]\) is the cylinder for the \(i\)-th sequence of length \(n\), \(C\subset X\) and \(L_n:\Sigma\to\mathcal M(\Sigma)\) gives the empirical measure in each point of \(\Sigma\). The dynamical multifractal zeta function is defined as \[ \varsigma U_C(z)=\sum\limits_{n=0}^{\infty}\frac{z^n}{n}\left(\sum\limits_{\overset{| i|=n}{UL_n[i]\subset C}}\sup_{u\in[i]}\exp\left(\sum\limits^{n-1}_{i=0}\varphi(\sigma^i(u))\right)\right). \] If \(C=\{\alpha\}\), \(\alpha\in\mathbb R\), then the above definitions capture the information about the usual multifractal decomposition. For instance if \(U\) is the functional \(U_\varphi(\mu)=\int\varphi d\mu\) then \(UL_n\) are the Birkhoff averages. The main goals are proving a variational principle for the multifractal pressure and solving the multifractal Bowen equation, i.e., finding a value \(F(C)\) for which the convergence radius of \(\varsigma\) is equal to 1. These result are established in a general way and apply to particular cases like Birkhoff decomposition and dimension spectra.