This article is cited in
10 papers
Well-posedness of approximation and optimization problems for weakly convex sets and functions
G. E. Ivanov,
M. S. Lopushanski Moscow Institute of Physics and Technology (State University), Moscow, Russia
Abstract:
We consider the class of weakly convex sets with respect to a quasiball in a Banach space. This class generalizes the classes of sets with positive reach, proximal smooth sets and prox-regular sets. We prove the well-posedness of the closest points problem of two sets, one of which is weakly convex with respect to a quasiball
$M$, and the other one is a summand of the quasiball
$-rM$, where
$r\in(0,1)$. We show that if a quasiball
$B$ is a summand of a quasiball
$M$, then a set that is weakly convex with respect to the quasiball
$M$ is also weakly convex with respect to the quasiball
$B$. We consider the class of weakly convex functions with respect to a given convex continuous function
$\gamma$ that consists of functions whose epigraphs are weakly convex sets with respect to the epigraph of
$\gamma$. We obtain a sufficient condition for the well-posedness of the infimal convolution problem, and also a sufficient condition for the existence, uniqueness, and continuous dependence on parameters of the minimizer.
UDC:
517.982.252