Local and semilocal Poincaré inequalities on metric spaces (Q1791738)

During the last two decades, analysis on metric spaces has seen big progress. Especially, the first-order theory has been investigated thoroughly. In Euclidean spaces, the fundamental theorem is very important to have control of a function on many lines. In metric measure spaces, one does not have lines but curves. However, in order to have a viable theory, one needs a lot of curves to ensure that the variants of the fundamental theorem can unfold their power. It turns out that the so called \textit{Poincaré inequalities} play an important role to guarantee enough curves. However, these inequalities are global, while conclusions like differentiability or the concept of Lebesgue points are of a more local character. This paper investigates local and semilocal Poincaré inequalities. In Section 3, results concerning doubling spaces and measures (the local and semilocal versions) are collected. Also, a local version of Lebesgue's differentiation theorem is established. The next section investigates the variants of Poincaré inequalities and connectivity. The main result is that under certain conditions, a local \((q,p)\)-Poincaré inequality implies a semilocal one. Following is the study of the improvement of the exponents of the Poincaré inequalities. There are two types of improvement; improving the exponent on the left-hand side leading to Sobolev-Poincaré inequalities (done in the global setting by \textit{P. Hajłasz} and \textit{P. Koskela} [C. R. Acad. Sci., Paris, Sér. I 320, No. 10, 1211--1215 (1995; Zbl 0837.46024); Mem. Am. Math. Soc. 688, 101 p. (2000; Zbl 0954.46022)]) and on the right-hand side (done in the global setting by \textit{S. Keith} and \textit{X. Zhong} [Ann. Math. (2) 167, No. 2, 575--599 (2008; Zbl 1180.46025)]). The authors of the article under review give local versions of these theorems. In the following section, the authors upgrade the results to semiuniform local assumptions. Next, it is shown under local assumptions that quasi every point of a function in a Newtonian Sobolev space is a Lebesgue point. The approach in this paper is novel as earlier approaches used the density of continuous functions in Newtonian spaces. Here, the authors use that Newtonian functions are more precisely defined than only almost everywhere. Although density of Lipschitz functions is not needed in above proofs, its investigation is still important. There are two different global results. One by \textit{N. Shanmugalingam} [Rev. Mat. Iberoam. 16, No. 2, 243--279 (2000; Zbl 0974.46038)] using the doubling property and the Poincaré inequality. The other approach is by \textit{L. Ambrosio} et al. [Rev. Mat. Iberoam. 29, No. 3, 969--996 (2013; Zbl 1287.46027)] using no Poincaré inequality, but still using that the metric space is doubling and complete instead. In Section 9, the authors use the density of Lipschitz functions to deduce results about quasicontinuity. The authors end the paper by applying the obtained results to \(p\)-harmonic functions.