Bounds for the \(L^q\)-spectra of self-similar measures without any separation conditions (Q645370)

Let \((S_1,\dots, S_N)\) be contracting similarities in \(\mathbb{R}^d\) and \({\mathbf p}:=(p_1,\dots, p_N)\) a probability vector. It follows from \textit{J. E. Hutchinson} [Indiana Univ. Math. J. 30, 713--747 (1981; Zbl 0598.28011)] that there exist a unique non-empty compact subset \(K\) of \(\mathbb{R}^d\) and a unique Borel probability measure \(\mu\) on \(\mathbb{R}^d\) supported by \(K\) such that \[ K=\bigcup_{i=1}^N S_i(K) \quad\text{and}\quad \mu= \sum_{i=1}^N p_i\mu \circ S_i^{-1}. \] The measure \(\mu\) is called the self-similar measure associated with the list \((S_1,\dots, S_N)\) and the vector \(\mathbf p\). Given \(q\in \mathbb{R}\), the lower and upper \(L^q\)-spectra of \(\mu\) are defined respectively by \[ \underline{\tau}(q):= \liminf_{r\to 0^+} \frac {\log I_\mu(r;q)}{-\log r} \quad \text{and} \quad \overline{\tau}(q):= \limsup_{r\to 0^+} \frac {\log I_\mu(r;q)}{-\log r}, \] where, for every \(r>0\), \[ I_\mu(r;q):= \int_K \mu(B(x,r))^{q-1} d\mu(x). \] \textit{M. Arbeiter} and \textit{N. Patzschke} [Math. Nachr. 181, 5--42 (1996; Zbl 0873.28003)] have provided a precise formula for the \(L^q\)-spectra of self-similar measures \(\mu\) associated with lists \((S_1,\dots,S_N)\) satisfying the Open Set Condition (OSC). Unfortunately, computing exactly the \(L^q\)-spectra of self-similar measures without OSC is hard. \textit{L. Olsen} [J. Math. Anal. Appl. 355, 12--21 (2009; Zbl 1170.28002)] has recently obtained non-trivial bounds of the \(L^q\)-spectra of arbitrary self-similar measures (without any separation condition) in the case \(q\geq 1\). In this paper the authors complete their investigation by giving non-trivial bounds of the \(L^q\)-spectra of arbitrary self-similar measures for \(q\leq 1\). This problem is more difficult than (and cannot be approached similarly to) Olsen's previous one, to make it clear in the latter case the \(L^q\)-spectra are extremely sensitive to small variations in the distribution of \(\mu\).