The restriction theorem for fully nonlinear subequations (Q486750)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | The restriction theorem for fully nonlinear subequations |
scientific article |
Statements
The restriction theorem for fully nonlinear subequations (English)
0 references
16 January 2015
0 references
Let \(J^2(\mathcal U)\) be the 2-jet space of real functions on an open subset \(\mathcal U \) of \( \mathbb R^N\), that is the collection of all triples \(j^2_x u = (u|_x, D_xu, D^2_x u)\), \(x \in \mathcal U\), given by the value \(u|_x\) of a \(\mathcal C^2\) real function \(u\) on a neighbourhood of \(x\), its first derivatives \(D_x u = (\partial_{x_i} u)|_x\) and the Hessian \(D^2_x u = (\partial^2_{x_i x_j} u)|_x\). A \textit{subequation on \(\mathcal U\)} is a closed subset \(F \subset J^2(\mathcal U)\) satisfying the following condition: if the 2-jet \(j^2_x u\) is in \(F\), then also the 2-jet \(j^2_x u + (0, 0, A)\) is in \(F\) for any positive semi-definite symmetric real matrix \(A = (A_{ij})\). Given a subequation \(F\), the \textit{\(F\)-subharmonic functions on \(\mathcal U\)} are the upper semi-continuous functions \(u:\mathcal U \to \mathbb R\) for which the following holds: for all \(x \in \mathcal U\) and all \(\mathcal C^2\)-functions \(\varphi\) with \(u(x) = \varphi(x)\) and \(u \leq \varphi\) on a neighbourhood of \(x\), the 2-jet \(j^2_x \varphi\) is in \(F\). First examples of \(F\)-subharmonic functions are provided by the plurisubharmonic functions and the convex functions in viscosity sense. Other important examples are the viscosity solutions to the Monge-Ampère equation \(\det(D^2 u) = 0\) with a fixed minimum number of non-negative eigenvalues for the Hessians. The main problem addressed in the paper is the following. Let \(F\) be a subequation on \(\mathcal U \subset \mathbb R^N\), \(N = n + m\), and set \(X := \{\;x^{n+1} = \ldots = x^N = 0\;\}\cap \mathcal U\). Denote by \(\imath^*: J^2(\mathcal U) \to J^2(X)\) the standard restriction map, that is \[ \imath^*\big((u, (\partial_{x_i} u), (\partial^2_{x_i x_j} u))\big|_x\big) = (u , (\partial_{x_i} u)_{i = 1, \ldots, n}, (\partial^2_{x_i x_j} u)_{i,j = 1, \ldots, n})\big|_x\;. \] Denoting \(F^{(X)} := \overline{\imath^*F}\), we say that a subequation \(F\) satisfies the \textit{restriction property to \(X\)} if \[ u \text{ is } F\text{-subharmonic on }\mathcal U\qquad \Longrightarrow \qquad u|_X \text{ is } F^{(X)}\text{-subharmonic on } \mathcal U. \] In case \(F\) is the subequation that characterises the plurisubharmonic functions of a complex manifold in a system of complex coordinates, the claim that \(F\) satisfies the restriction property for an appropriate \(X\) yields the property that a function is plurisubharmonic in viscosity sense if and only if all its restrictions to complex curves are subharmonic. In fact, one of the main results of this paper on the restriction property has been recently used by the same authors to prove the corresponding statement on almost complex manifolds [\textit{F. R. Harvey} and \textit{H. B. Lawson jun.}, ``Potential theory on almost complex manifolds'', Ann. Inst. Fourier (to appear), \url{arXiv:1107.2584}]. In the present paper the authors investigate in full generality the conditions under which a subequation \(F\) on an open subset \(\mathcal U\) satisfies the restriction property. They provide examples where this does not occur and they single out a simple condition, called \textit{restriction hypothesis}, that implies the restriction property. Then they exhibit a number of subequations that satisfy the restriction hypothesis and obtain important applications in many different contexts as, e.g., potential theory for calibrated geometries and universal subequations in Riemannian geometry.
0 references
viscosity solution
0 references
viscosity subsolution
0 references
nonlinear second-order elliptic equations
0 references
pluripotential theory
0 references