Nonconvexity in problems of multivalued calculus (Q1585389): Difference between revisions

Let \(P\) be a nonempty closed subset of a normed space \(X\). The function \[ \alpha_P(r):= \sup\sup\{r^{-1}\text{dist}(q, P): q\in \text{Conv}(P\cap K(x,r))\}, \] where the outer sup is taken over all open balls \(K(x,r)\) intersecting \(P\) and whose radius is \(r> 0\), is called the nonconvexity function of \(P\). A set \(P\) is called \(\alpha\)-paraconvex if its nonconvexity function does not exceed (pointwise) the function \(\alpha\). By using the concepts of this kind, there are surveyed the recent results concerning: extensions of continuous mappings, continuous and approximate selections, \(\varepsilon\)-approximations, fixed-point and minimax theorems for multifunctions. A nonconvexity function is applied also to some generalized convexities on metric spaces. The bibliography contains 55 positions, but is still far to be complete. Remark: The English version contains some mistakes deforming the meaning of the text, e.g.: \(2683_1\) nondegeneracy -- should be: degeneracy, \(2688^{21}\) \dots sets \textit{of} the union -- \textit{or} the union\dots, \(2699_1\) reference 55 is without the name of the journal, etc.