Sufficient optimality conditions for discrete minmax problems in the presence of constraints in Banach spaces (Q2277150)

We consider the following discrete minimax problem: \[ (1)\quad Minimize\quad \max f_ i(x),\quad i\in [0,...,N],\quad s.t.\quad F(x)\in K\text{ and } x\in C, \] where \(f_ i\) is a functional defined on a real Banach space X \((i=0,1,...,N)\), C is a non-empty closed convex subset of X, F is a map from X into a real Banach space Y and K is a closed convex cone in Y with vertex at the origin. The discrete minimax problem (1) is a nonsmooth problem which is closely connected with the smooth one studied by \textit{J. Zowe} and \textit{S. Kurcyusz} [Appl. Math. Optimization 5, 49-62 (1979; Zbl 0401.90104)] and \textit{V. F. Dem'yanov} and \textit{V. N. Malozemov} [``Introduction to minimax'' (Russian) (1972; Zbl 0245.90035)]. In the finite-dimensional case, necessary and sufficient optimality conditions for problem (1) involving only the second constraint were given by the author [Acta Math. Vietnam. 13, No.2, 87-98 (1988; Zbl 0713.49032)]. In the case of infinite-dimensional spaces necessary conditions for problem (1) with \(N=0\) were established in the paper of Zowe and Kurcyucz (loc. cit.), and sufficient conditions were studied in the book of Dem'yanov and Malozemov (loc. cit.). This paper presents some results concerning sufficient optimality conditions for problem (1). The paper consists of 4 sections. After the introduction, in Sections 2 and 3, using an approximation property of the feasible set, we obtain first and second-order sufficient optimality conditions for problem (1). Section 4 is devoted to the discussion of first-order sufficient optimality conditions for the problem with inequality-type constraint involving a finite number of functionals. The results of the paper yield in particular some known results including those of the above-cited works.