For a probability distribution $P$, Renyi fractal dimension is defined as

$$D_q = \lim_{\epsilon\rightarrow 0} \frac{R_q(P_\epsilon)}{\log(1/\epsilon)},$$ where $R_q$ is Renyi entropy of order $q$ and $P_\epsilon$ is the coarse-grained probability distribution (i.e. put in boxes of linear size $\epsilon$).

The question is if there are any phenomena, for which using non-trivial $q$ (i.e. $q\neq0,1,2,\infty$) is beneficial or naturally preferred?


The Rényi entropy of order $q = \frac{1}{2}$ apperas in several measures of pure states entanglement, please see for example: Karol Zyczkowski, Ingemar Bengtsson: Relatively Pure states entanglement. This entropy has the property that for three state systems, the equientropy trajectories form circles with respect to the the Bhattacharyya distance, please see for example: Bengtsson Zyczkowski: Geometry of quantum states, page 57.

  • $\begingroup$ Thank you for a nice example. However, $q=1/2$ is semi-trivial (as it is conjugated to $q=\infty$). $\endgroup$ – Piotr Migdal Dec 16 '11 at 10:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.