Viscosity solutions of fully nonlinear second-order elliptic partial differential equations (Q920285): Difference between revisions

This is a long and technical paper about viscosity solutions for fully nonlinear elliptic equations \(F(x,u,\nabla u,D^ 2u)=0\) in \(\Omega\) under various boundary conditions. The common strategy to tackle these equations is the observation of the first author [Duke Math. J. 55, 362- 384 (1987) and Commun. Pure Appl. Math. 42, No., 15-45 (1989; Zbl 0645.35025)] that (unique) existence is implied by a Perron-process, if viscosity sub- and supersolutions are known and a kind of maximum principle can be proved. It reads as follows: Whenever u (resp. v) is an usc (resp. lsc) bounded viscosity sub- (resp. super-) solution, then \[ u-v\leq \sup_{x\in \partial \Omega}\{u^*(x)-v_*(x)\}^+\text{ in } \Omega \] \(\sup_{y\in \Omega,y\to x} \sup u(y)\) and \(v_*=-(-v)^*.\) Hence the problem remains in (and most of the paper is devoted to) veryfying this under various structure conditions on F, including Isaac- Bellman equations and also Monge-Ampère equations. See also \textit{R. Jensen} [Arch. Ration. Mech. Anal. 101, No.1, 1-27 (1988)]. The paper closes with some remarks to the regularity of solutions. For \(C_{1,\alpha}\)-estimates, see also \textit{N. S. Trudinger} [Proc. R. Soc. Edinb. Sect. A 108, No.1/2, 57-65 (1988; Zbl 0653.35026)].