Second-order optimality conditions for inequality constrained problems with locally Lipschitz data (Q601984): Difference between revisions

Necessary and sufficient conditions for a point to be a local minimizer and an isolated local one of order two (see e.g., [\textit{A. Auslender}, ``Stability in mathematical programming with nondifferentiable data'', SIAM J. Control Optimization 22, 239--254 (1984; Zbl 0538.49020)]) are presented for inequality constrained problems in non-smooth settings with locally Lipschitz data. Second-order optimality conditions are of Fritz John and Karush-Kuhn-Tucker type. In the primal necessary conditions all functions are locally Lipschitz and in all other conditions the data are locally Lipschitz, regular in the sense of Clarke, Gâteaux differentiable at the mentioned point, and the constraint functions are second-order Hadamard differentiable at the point in every direction in the sense introduced in the present paper. It is shown by an example that regularity and Gâteaux differentiability cannot be removed from the sufficient conditions. Second order optimality conditions for inequality constrained problems with \(C^1\) data in the previous papers [\textit{I. Ginchev} and \textit{V. I. Ivanov}, ``Second order optimality conditions for problems with \(C^1\) data'', J. Math. Anal. Appl. 340, No. 1, 646--657 (2008; Zbl 1190.90208); \textit{V. I. Ivanov}, ``Optimality conditions for an isolated minimum of order two in \(C^1\) constrained optimization'', J. Math. Anal. Appl. 356, No. 1, 30--41 (2009; Zbl 1176.90569)] are extended to problems with locally Lipschitz and Gâteaux differentiable objective and constraint functions.