Abstract
We consider the problem (Newton's problem) and its generalizations. In the paper by Brock, Ferone, and Kawohl (1996) it is proved that if a solution u is C2 in an open set then det D2u = 0 in . It follows that graph does not contain extreme points of the subgraph of u. In this paper we prove a somewhat stronger result. Namely, there exists a solution u possessing the following property. If u is C1 in an open set then graph does not contain extreme points of the convex body Cu = {(x, z): x ∈ Ω, 0 ⩽ z ⩽ u(x)}. As a consequence, we have , where SingCu denotes the set of singular points of ∂Cu. We prove a similar result for a generalization of Newton's problem.
Export citation and abstract BibTeX RIS
Recommended by Professor Dmitry V Treschev
Footnotes
- 1
Actually, Newton considered a one-parameter family of reflection laws. Namely, a parameter 0 ⩽ k ⩽ 1 is fixed, and in a reference system connected with the body, at each impact, the normal component of the particle's velocity of incidence is multiplied by −k, while the tangential component remains unchanged. In the case k = 1 we have the law of perfectly elastic (billiard) reflection.
- 2
A convex body is a compact convex set with nonempty interior.
- 3
- 4
A point ξ ∈ ∂Ω is called regular, if there is a unique line of support to Ω at ξ.
- 5
The authors work with convex, rather than concave, functions u1, u2, u satisfying u2 ⩽ u ⩽ u1. With these changes, the statement remains the same.
- 6
A point (x, u(x)) of the graph is singular, if u is not differentiable at x.