Intersecting sets in midset spaces. I (Q1842016): Difference between revisions
From MaRDI portal
Added link to MaRDI item. |
Set profile property. |
||
Property / MaRDI profile type | |||
Property / MaRDI profile type: MaRDI publication profile / rank | |||
Normal rank |
Revision as of 04:54, 5 March 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Intersecting sets in midset spaces. I |
scientific article |
Statements
Intersecting sets in midset spaces. I (English)
0 references
18 April 1995
0 references
In this paper, divided in two parts but reviewed together, the author continues work begun by the author and \textit{R. Trost} [Acta Math. Hung. 54, No. 1/2, 39-49 (1989; Zbl 0688.90061)] and the author [Proc. Am. Math. Soc. 117, No. 4, 1003-1011 (1993; Zbl 0774.54016) and J. Math. Anal. Appl. 178, No. 2, 529-546 (1993; Zbl 0786.49007)]. The following is a typical result of Part I: Let \(Y\) be a nonvoid set, \(X\) an index set, and \(\{C_ x: x\in X\}\) a system of nonvoid subsets of \(Y\). Then the intersection \(\bigcap\{C_ x: x\in X\}\) is nonvoid iff \(X\) and \(Y\) can be endowed with topologies and with ``midset'' functions \(Z_ X: X\times X\to 2^ X\) and \(Z_ Y: Y\times Y\to 2^ Y\) such that (i) \(Y\) is compact, (ii) the midsets \(Z_ X(x_ 1, x_ 2)\supset \{x_ 1, x_ 2\}\), \((x_ 1, x_ 2)\in X\times X\) and \(Z_ Y(y_ 1, y_ 2)\supset \{y_ 1, y_ 2\}\), \((y_ 1, y_ 2)\in Y\times Y\), are connected, (iii) every set \(C_ x\), \(x\in X\), is \(Z_ Y\)-convex (i.e., \(\{y_ 1, y_ 2\}\subset C_ x\) implies \(Z_ Y(y_ 1, y_ 2)\subset C_ x\)), (iv) every ``dual'' set \(C^*_ y:= \{x\in X: y\not\in C_ x\}\) is \(Z_ X\)-convex, (v) every set \(C_ x\) is closed, and (vi) \(\bigcap \{C^*_ y: y\in F\}\) is open for every closed \(F\subset Y\) (i.e., the correspondence \(x\to C_ x\) is upper semicontinuous). In Part II (see below) several applications are given. For example, the following Markov-Kakutani type fixed point theorem is proved: Let \((Y, d)\) be a compact, Menger-convex metrix space (i.e., \((y_ 1, y_ 2):= \{y\in Y: d(y_ 1, y)+ d(y, y_ 2)= d(y_ 1, y_ 2)\}- \{y_ 1, y_ 2\}\) is nonvoid for \(y_ 1\neq y_ 2\)). Suppose that the set \(\mathcal A\) of all continuous \(d\)-affine functions \(f: Y\to \mathbb{R}\) (characterized by \(f(y_ 0) d(y_ 1, y_ 2)= f(y_ 1) d(y_ 2, y_ 0)+ f(y_ 2) d(y_ 0, y_ 1),y_ 0\in (y_ 1, y_ 2))\) separates points. Then every family \({\mathcal T}\) of continuous, pairwise commuting functions \(T: Y\to Y\) with \(f\circ T\in {\mathcal A}\) for all \(f\in {\mathcal A}\), \(T\in {\mathcal T}\), possesses a common fixed point. In addition, intersection theorems are proved where only topological properties are involved. As applications generalized versions of minimax theorems due to Ha, Jóo, Komiya, Komornik, König, Simons, Sion, Stachó, and Tuy are obtained.
0 references
intersecting sets
0 references
connectedness
0 references
convexity
0 references
quartercontinuous correspondences
0 references
Markov-Kakutani type fixed point theorem
0 references
intersection theorems
0 references
minimax theorems
0 references