No label defined (Q4536112)

This nice piece of thesis work continues the study of the geometry of the average dimension \(\dim_{A}\), started in [\textit{M. Zähle}, Fractals 3, No. 4, 747-754 (1995; Zbl 0866.28007)], where \(\dim_A\) was defined from the notion of order-two density \(d_A^\alpha\mu(x)\) of \textit{T. Bedford} and \textit{A. M. Fisher} [Proc. Lond. Math. Soc., III. Ser. 64, No. 1, 95-124 (1992; Zbl 0706.28009)]. The local average dimension of a measure \(\mu\) at \(x\in{\mathbb R}^n\) is given by \[ \dim_{A}\mu(x)=\sup\left\{\alpha\geq 0: d_{A}^{\alpha}\mu(x):=\liminf_{\delta\rightarrow 0}{1 \over |\log\delta|}\int_{\delta}^1 {\mu B(x,r) \over r^\alpha} {1 \over r}dr =0 \right\}. \] Natural geometric measure-theoretic properties of \(\dim_A\) are considered for compactly supported Radon probability measures, namely, behaviour for slice measures, product measures, and intersection measures. It is proved that \(\dim_A\) behaves like the Hausdorff dimension under slicing. Sharp versions of the known (Hausdorff and packing) dimension formulae for product measures are derived for the average dimension. Elucidating examples are supplied showing that the bounds for the Hausdorff and packing dimensions of product measures -- in terms of Hausdorff and packing measures of the factor measures -- cannot be improved using their average dimensions instead (that lies in between in general). Lastly, analogues of the results for intersection measures in [\textit{M. Järvenpää}, Ann. Acad. Sci. Fenn., Math. 24, No. 1, 165-186 (1999; Zbl 0914.28004)] are obtained for the average dimension.