The Calabi conjecture on affine maximal surfaces (Q5954157): Difference between revisions

This paper gives an affirmative answer to a famous conjecture of \textit{E. Calabi} in Am. J. Math. 104, 91-126 (1984; Zbl 0501.53037). Consider a locally strongly convex surface in real affine 3-space. In the conformal class of the Euclidean second fundamental form there is exactly one Riemannian metric which is invariant under the unimodular transformation group of the ambient space; this metric nowadays is called Blaschke metric. The variational problem for its area functional leads to an Euler-Lagrange equation, stating that the affine mean curvature identically vanishes. As the second variation of the area functional is negative, Calabi called this class of surfaces affine maximal. Chern and Calabi stated two different versions of a so called affine Bernstein conjecture, differing in the completeness assumptions. Chern's conjecture from 1978 (every global graph over \(R^2\) which is affine maximal is an elliptic paraboloid) has been answered affirmatively by \textit{N. Trudinger} and \textit{X. J. Wang} in Invent. Math. 140, No. 2, 399-422 (2000; Zbl 0978.53021). The present authors prove Calabi's conjecture from 1982 (every affine maximal surface which is complete with respect to the Blaschke metric is an elliptic paraboloid). Using a Lemma of \textit{H. Hofer} in Invent. Math. 114, No. 3, 515-563 (1993; Zbl 0797.58023) on positive continuous functions on complete metric spaces, the authors show that the norm of the affine Weingarten operator of the affine maximal surface is globally bounded from above. This now allows to apply a result of \textit{A. Martínez} and \textit{F. Milán} [Geom. Dedicata 37, No. 3, 295-302 (1991; Zbl 0715.53011)] to give the solution of the affine Calabi conjecture. The idea of proof is new and highly interesting, it might be that one can extend it to the investigation of metrically complete Blaschke surfaces with constant negative affine mean curvature.