O problema de Bernstein

Abstract

The classical Bernstein problem, solved by S. Bernstein in 1915-1917 in his article [12], asks if there is a complete minimal graph in R3 besides the plane. Bernstein showed that the answer to this question is no using analytical methods for study of equations of prescribed curvature. We will see here how this problem is related to the Gauss map of the graph, and as consequence of this relationship we generalize this theorem to a larger class of surfaces (not necessarily graphs), following the proof given by R. Osserman in [51]. We will see next generalizations of this theorem in higher dimensions, following essentially the methods introduced by W. Fleming in [31], and later refined by E. De Giorgi in [20], F. Almgren in [6] and J. Simons in [62]. In fact, they solve the problem for graphs in Rn, n < 9, namely they prove that the only complete minimal graph in these espaces is the hyperplane. Following the proof given by E. Bombieri, E. De Giorgi and E. Giusti in [14], we also show that, in dimension n ≥ 9, it is possible to construct complete minimal graphs in Rn. At last, we conclude with an extension of Bernstein’s theorem to the class of submanifolds stable with respect to the second variation of volume, under certain conditions of curvature and volume growth, and yet we investigate the case in which the ambient manifold is not the Euclidean space.