A variational principle for fractal dimensions (Q2492394): Difference between revisions

The author introduces the notion of the so-called thin dimension of a probability measure on a separable metric space. The author goes on to show that upper bounds for the upper (resp. lower) fractal dimension of any subset in the support of such a probability measure can be obtained by calculating the upper (resp. lower) thin dimension of the measure. This result is then sharpened somewhat to obtain precise upper bounds for the fractal dimension of a non-empty compact subset of the underlying metric space. Using a variational principle, a probability measure is constructed whose support is exactly that of the compact subset and further, the thin dimension of the measure coincides with the upper fractal dimension. The paper concludes with an application of the theory to condensation sets, a well-known concept in the field of iterated function systems. The author establishes upper bounds for both the fractal and thin dimensions depending only on the associated Lipshitz constants and probability vectors in the condensation systems.