KLS-type isoperimetric bounds for log-concave probability measures (Q292624): Difference between revisions

Let \(\mu\) be a log-concave probability measure on \(\mathbb{R}^n\) with density \(f\) and let \(A\) be an arbitrary Borel subset of \(\mathbb{R}^n\) of measure \(\mu(A)\) with \(\mu\)-perimeter \[ \mu^+(A) = \liminf_{\varepsilon \rightarrow 0} \frac{\mu(A_{\varepsilon}) - \mu(A)}{\varepsilon}, \] in which \(A_{\varepsilon}\) denotes an \(\varepsilon\)-neighborhood of \(A\). The authors study the optimal value \(h=h(\mu)\) in the isoperimetric-type inequality \[ \mu^+(A) \geq h \min \{\mu(A), 1-\mu(A) \}. \] \textit{R. Kannan} et al. [Discrete Comput. Geom. 13, No. 3--4, 541--559 (1995; Zbl 0824.52012)] studied the particular case in which \(\mu=\mu_K\) is the uniform distribution in a given convex body \(K\) in \(\mathbb{R}^n\). The main result of the paper under review extends this result to general log-concave probability measures; in particular the authors prove an upper bound for \(h(\mu)^{-1}\) for any probability measure \(\mu\) on \(\mathbb{R}^n\) with a log-concave density \(f\). This interesting and general result is illustrated in Section 5 with 8 different examples. The authors also point out that a variant of their main result, Corollary 1.2, unites two different bounds for \(h(\mu)\) that were previously thought to be non-comparable.