Fix $c>1$. Let $(X,d)$ be a separable compact metric space, does there necessarily exist a Borel probability measure $\nu$ on $(X,d)$ such that

  • $\operatorname{sup}_{x \in X,r>0}\frac{\nu(\mathrm{Ball}(x,cr))}{\nu(\mathrm{Ball}(x,r))}<\infty$,
  • $\nu(\mathrm{Ball}(x,r))>0$ for every $x \in X$ and $r>0$.
  • 1
    $\begingroup$ What is $c$? Also, why doesn't the second point imply the first? Or do you want some uniform boundedness? $\endgroup$ – Steve Jun 26 '19 at 15:56
  • $\begingroup$ I clarified c and point 1. It isn't implied by the second one anymore (excuse my typo) $\endgroup$ – AIM_BLB Jun 26 '19 at 19:27

Are you asking if there is a $c>1$ and a measure $\nu$ with this property? And is the first condition supposed to hold with a uniform upper bound? (If not, any probability measure satisfying the second condition would work, as Steve points out.)

If so, you are asking for the existence of a non-trivial "doubling measure". A necessary and sufficient condition for a complete metric space to carry a non-trivial doubling measure is that it is a "doubling metric space".

A metric space is a "doubling metric space" if every ball $B(x,r)$ can be covered by $N$ balls of radius $r/2$, where $N$ is a fixed constant. See wikipedia: https://en.wikipedia.org/wiki/Doubling_space

If a complete metric space carries a non-trivial doubling measure, it must be a doubling metric space, by a fairly standard covering argument. (See, for example, Heinonen's book Lectures on Analysis on Metric Spaces.)

The reverse implication, that every complete doubling metric space carries a non-trivial doubling measure, is more difficult. It was proven by Vol'berg-Konyagin in the compact case and Luukkainen-Saksman in the general case. See Luukkainen-Saksman (https://www.ams.org/journals/proc/1998-126-02/S0002-9939-98-04201-4/S0002-9939-98-04201-4.pdf) for the details and references.

| cite | improve this answer | |
  • $\begingroup$ Doubling property is definitely sufficient but I'm not sure it's necessary. Here c is given and fixed but I'm not requiring that the balls all be compared by the same doubling constant. For example, I think this should hold for the unit ball in $\ell^2$ which isn't doubling $\endgroup$ – AIM_BLB Jun 26 '19 at 19:26
  • 1
    $\begingroup$ @AIM_BLB The metric doubling condition is necessary. See Proposition 59 in Measure Theory that you can find here: sites.google.com/view/piotr-hajasz/teaching/… $\endgroup$ – Piotr Hajlasz Jun 26 '19 at 20:19
  • $\begingroup$ Hi Piotr, thanks I worked it out on the Tram home and indeed it is necessary. I just found it surprising since the paper of Ambrosio which it comes from treated the two cases separately. $\endgroup$ – AIM_BLB Jun 26 '19 at 20:32

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.