Hölder continuous solutions to Monge-Ampère equations (Q2449821)

Let \((X,\omega)\) be a compact Kähler manifold of dimension \(n\). The authors study the following complex Monge-Ampère operator \[ \text{MA}(u):=\frac {1}{V_{\omega}}(\omega+dd^cu)^n, \;\;\text{ where } \;\;V_{\omega}=\int_X\omega^n, \tag{MA} \] acting on \(\omega\)-plurisubharmonic functions, \(u\in \mathrm{PSH}(X,\omega)\), which are Hölder continuous, \(u\in \text{Hölder}(X,\mathbb R)\). The following theorem gives a better information about the Hölder exponent of the solution to the complex Monge-Ampère equation than theorems proved recently by \textit{P. Eyssidieux} et al. [J. Am. Math. Soc. 22, No. 3, 607--639 (2009; Zbl 1215.32017)] and \textit{S. Dinew} [J. Inst. Math. Jussieu 9, No. 4, 705--718 (2010; Zbl 1207.32034)]. { Theorem~A.} Let \(\mu=f\omega^n=\text{MA}(u)\) be a probability measure absolutely continuous with respect to the Lebesgue measure with density \(f\in L^p\), \(p>1\). Then \(u\) is Hölder continuous with exponent arbitrary close to \(\frac {2}{1+nq}\), where \(\frac {1}{p}+\frac {1}{q}=1\). The optimal value of the Hölder exponent in Theorem A is still unknown, but it cannot be better than \(\frac {2}{nq}\), see [\textit{S. Pliś}, Ann. Pol. Math. 86, No. 2, 171--175 (2005; Zbl 1136.32306)] or [\textit{V. Guedj} et al., Bull. Lond. Math. Soc. 40, No. 6, 1070--1080 (2008; Zbl 1157.32033)]. Moreover, Theorem A is generalized from the Kähler case to the case of big cohomology classes. The rest of the paper is devoted to the study of the range \[ \text{MAH}(X,\omega)=\text{MA}\big(\text{PSH}(X,\omega)\cap \text{Hölder}(X,\mathbb R)\big). \] A complete characterization of the set \(\text{MAH}(X,\omega)\) is unknown, but some of its properties are proved in the following theorem. { Theorem~B.} The set \(\text{MAH}(X,\omega)\) has the \(L^p\) property, i.e. if \(\mu \in \text{MAH}(X,\omega)\), \(f\geq 0\), \(f\in L^p(\mu)\) with \(||f||_p=1\), then \(f\mu\in \text{MAH}(X,\omega)\). In particular \(\text{MAH}(X,\omega)\) is a convex set. \textit{T.-C. Dinh} et al. [J. Differ. Geom. 84, No. 3, 465--488 (2010; Zbl 1211.32021)] observed that the measures \(\mu\in \text{MAH}(X,\omega)\) have the following property \[ \text{exp}\big(-\epsilon \text{PSH}(X,\omega)\big)\subset L^1(\mu), \;\;\text{for some} \;\;\epsilon >0. \tag{DNS} \] Condition (DNS) gives a full description of the range \(\text{MAH}(X,\omega)\) for \(n=1\), see [\textit{T.-C. Dinh} and \textit{N. Sibony}, Comment. Math. Helv. 81, No. 1, 221--258 (2006; Zbl 1094.32005)]. In the article under review it is proved that, for some special class of measures, this characterization is true in higher dimensions. { Theorem~C.} Let \(\mu\) be a probability measure with finitely many isolated singularities which is radial or toric. Then \(\mu \in \text{MAH}(X,\omega)\) if and only if condition (DNS) is satisfied.