Generic well-posedness of minimization problems with mixed continuous constraints. (Q2490777): Difference between revisions

Let \(X,Y,Z\) be Banach spaces, with \(Y\) finite dimensional and \(Z\) ordered by a closed convex cone. The paper is concerned with the following mathematical programming problem with constraints: \[ \text{minimize}f(x)\text{ subject to }G(x) = y\text{ and }H(x)\leq z \] where \(f:X\to \mathbb R, \, G:X\to Y,\, H:X\to Z\) are continuous mappings and \(y\in Y, z\in Z.\) It is shown that for two classes of this kind of problems, most of the problems (in Baire category sense) are well-posed. These results extend those obtained by \textit{A. D. Ioffe} and the author [SIAM J. Control Optimization 38, No. 2, 566--581 (2000; Zbl 0997.49023)], for finite dimensional spaces \(Y\) and \(Z.\) The proofs are based on a general variational principle involving some minimization problems depending on a parameter ranging in a complete metric space, proved by the author in [SIAM J. Control Optimization 39, No. 1, 250--280 (2000; Zbl 0973.49022)].