Weak maximum principle for biharmonic equations in quasiconvex Lipschitz domains (Q2204360)

The validity of the weak maximum principle \[\|\nabla u\|_{L^\infty(\Omega)}\leq C\|\nabla u\|_{L^\infty(\partial \Omega)}\] for the biharmonic equation \(\Delta^2u=0\) in a bounded Lipschitz domain \(\Omega\subset \mathbb{R}^d\) is investigated. It is known that in dimension \(2\) or \(3\), the weak maximum principle for the biharmonic equation holds with no additional conditions on \(\Omega\), while it holds if \(\Omega\) is convex or \(C^1\) in dimension greater than or equal to 4. In this latter case, it is also known that the weak maximum principle fails if \(\Omega\) contains the exterior of a cone with small aperture. Here, the author proves that the weak maximum principle for the biharmonic equation is still valid under a more general quasi-convexity condition which is satisfied both by convex and \(C^1\) domains and allows domains containing the exterior of cones with sufficiently large aperture (thus, in this sense, it is a sharp condition). In particular, the author gives the following notion of quasi-convexity: given \(\delta,R>0\) and \(\sigma\in (0,1)\), a Lipschitz domain \(\Omega\) is said \((\delta,\sigma,R)\)-quasiconvex if, for any \(r\in (0,R)\) and \(Q\in \partial \Omega\), one has \[|\Omega\cap B_r(Q)|\geq \sigma |B_r(Q)\] and there exists a convex domain \(V=V(Q,r)\) such that \[\Omega\cap B_r(Q)\subset V \ \ \ \text{and} \ \ \ d(\partial (\Omega\cap B_r(Q)),\partial V):=\sup_{x\in \partial (\Omega\cap B_r(Q))}\inf_{y\in \partial V}|x-y|\leq \delta r.\] Then, he proves that there exists \(\delta_0>0\) depending only on the dimension \(d\) and the Lipschitz constant of \(\Omega\) such that the weak maximum principle for the biharmonic equation is valid if \(\Omega\) is \((\delta,\sigma,R)\)-quasiconvex with \(\delta \in (0,\delta_0)\). The main ingredients of the proof are: Meyer's estimate for biharmonic functions, the reverse Hölder inequality, and Shen's real variable method.