English

Mutually nearest and mutually farthest points of sets in geodesic spaces

Metric Geometry 2010-03-23 v1 Functional Analysis

Abstract

Let AA and XX be nonempty, bounded and closed subsets of a geodesic metric space (E,d)(E,d). The minimization (resp. maximization) problem denoted by min(A,X)\min(A,X) (resp. max(A,X)\max(A,X)) consists in finding (a0,x0)A×X(a_0,x_0) \in A \times X such that d(a0,x0)=inf{d(a,x):aA,xX}d(a_0,x_0) = \inf\{d(a,x) : a \in A, x \in X\} (resp. d(a0,x0)=sup{d(a,x):aA,xX}d(a_0,x_0) = \sup\{d(a,x) : a \in A, x \in X\}). We study the well-posedness of these problems in different geodesic spaces considering the set AA fixed. Let Pb,cl,cv(E)P_{b,cl,cv}(E) be the space of all nonempty, bounded, closed and convex subsets of EE endowed with the Pompeiu-Hausdorff distance. We show that in a space with a convex metric, curvature bounded below and the geodesic extension property, the family of sets in Pb,cl,cv(E)P_{b,cl,cv}(E) for which max(A,X)\max(A,X) is well-posed is a dense GδG_\delta-set in Pb,cl,cv(E)P_{b,cl,cv}(E). We give a similar result for min(A,X)\min(A,X) without needing the geodesic extension property. Besides, we analyze the situations when one set or both sets are compact and prove some results specific to CAT(0)(0) spaces. We also prove a variant of the Drop theorem in geodesic spaces with a convex metric and apply it to obtain an optimization result for convex functions.

Keywords

Cite

@article{arxiv.1003.4179,
  title  = {Mutually nearest and mutually farthest points of sets in geodesic spaces},
  author = {Rafa Espinola and Adriana Nicolae},
  journal= {arXiv preprint arXiv:1003.4179},
  year   = {2010}
}

Comments

20 pages, research paper