Alternative theorems and minimax inequalities in \(G\)-convex spaces (Q885319)

A generalized convex space or a \(G\)-convex space \((X,D;\Gamma)\) [\textit{S. Park}, Nonlinear Anal., Theory Methods Appl. 48, No. 6(A), 869--879 (2002; Zbl 1011.47048)] consists of a topological space \(X\) and a nonempty set \(D\) such that for each \(A\in\langle D\rangle\) with the cardinality \(|A|=n+1\) there exists a subset \(\Gamma(A)\) of \(X\) and a function \(\phi_A:\Delta_n\to \Gamma(A)\) such that \(J\in\langle A\rangle\) implies \(\phi_A(\Delta_J)\subset \Gamma(J)\). Here \(\langle D\rangle\) denotes the set of all nonempty finite subsets of \(D\), \(\Delta_n\) denotes any \(n\)-simplex with vertices \(\{e_i\}^n_{i=0}\) and \(\Delta\) the face of \(\Delta_n\) corresponding to \(J\). Suppose that \((X,D;\Gamma)\) is a \(G\)-convex space, with \(X\) Hausdorff, \(Y\) a non-empty set and \(T: X\multimap Y\), \(S: D\multimap Y\) two set-valued mappings. We say that \(S\) is a generalized \(G\)-KKM mapping w.r.t. \(T\) if for each \(A\in\langle D\rangle,T(\Gamma(A))\subset S(A)\). For a set-valued mapping \(S: X\multimap D\), we denote by \(\text{co}_\Gamma S: X\multimap Y\) the mapping defined by \(\text{co}_\Gamma S(x)= \bigcup\{\Gamma(A): A\in \langle S(x)\rangle\}\). The main theorem of the paper is Theorem 2: Let \((X,D_1;\Gamma_1)\) and \((Y,D_2;\Gamma_2)\) be \(G\)-convex spaces, one of them compact. Let \(T: X\multimap Y\), \(S_1: X\multimap D_2\) and \(S_2: D_1\multimap Y\) be three set-valued mappings satisfying the following conditions: (i) for each \(x\in X\), \(\text{co}_{\Gamma_2} S_1(x)\subset T(x)\); (ii) \(S_1\) has open fibers; (iii) \(S_2\) is a generalized \(G\)-KKM mapping w.r.t. \(T\); (iv) \(S_2\) has closed values. Then at least one of the following assertion holds: (a) there exists \(x_0\in X\) such that \(S_1(x_0)= \emptyset\); (b) \(\bigcap_{u\in D_1} S_2(u)\neq\emptyset\). Theorems 3, 4, 5 of the paper are equivalent to Theorem 2. As consequences, new alternative theorems and two very general minimax inequalities are obtained.