A Dirichlet problem for the complex Monge-Ampère operator in \(\mathcal F(f)\) (Q2460989): Difference between revisions

Let \(\Omega\subset\mathbb{C}^{n}\) be a bounded hyperconvex domain. The class of plurisubharmonic functions \(\mathcal{E}_{0}(\Omega)\) consists of all functions \(u\in \text{PSH} (\Omega)\cap L_{\infty }(\Omega) \) satisfying \(\lim_{z\rightarrow\xi}u(z) =0\) for every \(\xi\in\partial\Omega\) and \(\int_{\Omega}(dd^{c}u) ^{n}<\infty\) (each \(\varphi\in\mathcal{E}_{0}(\Omega) \) is either negative or zero). The complex Monge-Ampère operator \((dd^{c}) ^{n}\) is well defined on \(\mathcal{E}_{0}(\Omega) \) which can be viewed as a substitute for the space of test functions \(C_{0}^{\infty}(\Omega) \). Consider all functions \(u\) such that there is a sequence \(\{ \varphi_{j}\} _{j=1}^{n}\searrow u\) (pointwise convergence on \(\Omega\)), \(\{ \varphi_{j}\} _{j=1} ^{n}\subset\mathcal{E}_{0}( \Omega) \), and the following conditions: (a) for \(p\geq1\), \(\sup_{j}\int_{\Omega}( -\varphi _{j}) ^{p}(dd^{c}\varphi_{j}) ^{n}<+\infty\) and (b) \(\sup_{j}\int_{\Omega}(dd^{c}\varphi_{j}) ^{n}<+\infty\). Then \(u\in\mathcal{E}_{p}(\Omega) \) if (a) is satisfied; \(u\in\mathcal{F}(\Omega) \), if (b) is satisfied and \(u\in\mathcal{F}_{p}(\Omega) \) if both (a) and (b) are satisfied. \textit{U. Cegrell} [Ann. Inst. Fourier 54, No.~1, 163--184 (2004; Zbl 1065.32020)] introduced the set \(\mathcal{E}\) containing both \(\mathcal{F}\) and \(\mathcal{E}_{p}\) and proved that the complex Monge-Ampère operator \(( dd^{c}) ^{n}\) is well defined on \(\mathcal{E}(\Omega) \). The author introduces the envelope \(U( \mu,f) \), where \(f:\partial\Omega\rightarrow\mathbb{R}\) is a continuous function and \(\mu\) is a non-negative measure defined on \(\Omega\). The envelope \(U( \mu,f) \) is defined as supremum of all functions \(w\in \text{PSH}(\Omega) \cap L_{\text{loc}}^{\infty}( \Omega) \) satisfying \((dd^{c}w) ^{n}\geq\mu\) and \(\lim\sup_{z\rightarrow\xi}w( z) \leq f(\xi) \) for every \(\xi\in\partial\Omega\). Let \(f\) be such that \(\lim_{z\rightarrow\xi}U( 0,f) (z) =f(\xi) \) for all boundary points \(\xi\). If \(K=\{ \mathcal{F},\mathcal{E}_{p},\mathcal{E}\} \), then \(u\in K( f) =\{ \mathcal{F}( f) ,\mathcal{E}_{p}( f) ,\mathcal{E}( f)\} \) if there is a function \(\varphi\in K\) such that \(U( 0,f) \geq u\geq\varphi+U( 0,f) \). The author generalizes \textit{U. Cegrell}'s result [Acta Math. 180, No.~2, 187--217 (1998; Zbl 0926.32042)] by proving that the complex Monge-Ampère operator \((dd^{c}) ^{n}\) is well defined on \(\mathcal{F}(f) ,\mathcal{E}_{p}(f) \) and even on the larger set \(\mathcal{E}(f) \). The author's main result (Theorem 3.4) deals with Dirichlet's problem for the complex Monge-Ampère operator \((dd^{c}) ^{n}\) which generalizes a theorem by Cegrell [loc. cit.]. Suppose that \(\mu\) is a non-negative measure with finite \(\mu(\Omega) \) and zero values on pluripolar subsets of \(\Omega\). The author proves that for every continuous function \(f:\partial\Omega\rightarrow\mathbb{R}\) satisfying \(\lim _{z\rightarrow\xi}U(0,f) ( z) =f(\xi) \), for every \(\xi\in\partial\Omega\), the complex Monge-Ampère equation \((dd^{c}u) ^{n}=\mu\) has a unique solution in \(\mathcal{F}( f) \). As a corollary, the author proves the following comparison principle (Corollary 3.6). Let \(\mathcal{F}^{a}(f) \) be the set of \(u\in\mathcal{F}( f) \) such that \(( dd^{c}u) ^{n}\) vanishes on all pluripolar subsets of \(\Omega\). Suppose that \(f,g:\partial\Omega\rightarrow\mathbb{R}\) are continuous and \(\lim_{z\rightarrow\xi}U( 0,f) =f( \xi) \), \(\lim_{z\rightarrow\xi}U( 0,g) =g( \xi) \) for every \(\xi\in\partial\Omega\) and \(f\leq g\). If \(u\in\mathcal{F}( f) \) and \(v\in\mathcal{F}^{a}t( g) \), where \(\int_{\Omega}( dd^{c}u) ^{n}<+\infty\) and \(( dd^{c}u) \geq( dd^{c}v) ^{n}\), then \(u\leq v\).