First-order Sobolev spaces, self-similar energies and energy measures on the Sierpiński carpet
Mathav Murugan,Ryosuke Shimizu
2023-11-29
Abstract:We construct and investigate $(1, p)$-Sobolev space, $p$-energy, and the corresponding $p$-energy measures on the planar Sierpiński carpet for all $p \in (1, \infty)$. Our method is based on the idea of Kusuoka and Zhou [Probab. Theory Related Fields $\textbf{93}$ (1992), no. 2, 169--196], where Brownian motion (the case $p = 2$) on self-similar sets including the planar Sierpiński carpet were constructed. Similar to this earlier work, we use a sequence of discrete graph approximations and the corresponding discrete $p$-energies to define the Sobolev space and $p$-energies. However, we need a new approach to ensure that our $(1, p)$-Sobolev space has a dense set of continuous functions when $p$ is less than the Ahlfors regular conformal dimension. The new ingredients are the use of Loewner type estimates on combinatorial modulus to obtain Poincaré inequality and elliptic Harnack inequality on a sequence of approximating graphs. An important feature of our Sobolev space is the self-similarity of our $p$-energy, which allows us to define corresponding $p$-energy measures on the planar Sierpiński carpet. We show that our Sobolev space can also be viewed as a Korevaar-Schoen type space. We apply our results to the attainment problem for Ahlfors regular conformal dimension of the Sierpiński carpet. In particular, we show that if the Ahlfors regular conformal dimension, say $\dim_{\mathrm{ARC}}$, is attained, then any optimal measure which attains $\dim_{\mathrm{ARC}}$ should be comparable with the $\dim_{\mathrm{ARC}}$-energy measure of some function in our $(1, \dim_{\mathrm{ARC}})$-Sobolev space up to a multiplicative constant. In this case, we also prove that the Newton-Sobolev space corresponding to any optimal measure and metric can be identified as our self-similar $(1, \dim_{\mathrm{ARC}})$-Sobolev space.
Metric Geometry,Analysis of PDEs,Functional Analysis,Probability