On sublevel set estimates and the Laplacian (Q2039093)

The author is interested in estimates on the volume of the sublevel sets of suitable convex functions. He starts by giving a short, simple proof of \textit{A. Carbery}'s sharp result [Contemp. Math. 505, 97--103 (2010; Zbl 1190.26016)] that for any positive, strictly convex function \(u: \mathbb{R}^n \to \mathbb{R}\) satisfying \(\det D^2 u \geq 1\), we have \[ |\{x \in \mathbb{R}^n: u(x) \leq s\}| \leq C(n) s^{\frac{n}{2}} \] for all \(s > 0\), for a dimensional constant \(C(n)\). His alternative proof also leads to (or inspires) a level set estimate for functions with bounded, positive Laplacian in a model two-dimensional case: if (the sufficiently smooth function) \(u: [0,1]^2 \to \mathbb{R}\) satisfies \(1 \leq \Delta u \leq c\) for some \(c \geq 1\), then for all \(\varepsilon > 0\), considered small but not necessarily so, \[ |\{x \in [0,1]^2 : |u(x)| \leq \varepsilon \}| \leq C(c)\left[\sqrt{\varepsilon}+ \varepsilon^{\alpha-\frac{1}{2}}\int_{[0,1]^2} \frac{|\nabla u|}{|u|^\alpha}\,dx\right] \] for any \(\alpha \geq 0\). Thus, if \(\Delta u \sim 1\), then we can generall expect the sublevel sets to have volume like \(\varepsilon^{\frac{1}{2}-}\). The proof of the latter theorem combines elementary but well-chosen pointwise estimates on \(u\), the coarea formula, and tools and estimates (on Brownian motion) from Stochastic Analysis. The paper, which has a somewhat exploratory flavour, is well written and well motivated.