PROPER AND ISOLATED EFFICIENCIES IN MULTIOBJECTIVE OPTIMIZATION PROBLEMS

Tóm tắt

We employ some advanced tools of variational analysis and generalised differentiation such as the nonsmooth version of Fermat's rule, the limiting sub-differential of maximum functions, and the sum rules for the Frechet and Mordukhovich sub-differentials to establish necessary conditions for (local) properly efficient solutions and (local) isolated minimizers of a multi-objective optimisation problem involving inequality and equality constraints. Sufficient conditions for the existence of such solutions are also supplied under assumptions of (local) convex/affine functions or L-invex-infine functions defined in terms of the limiting sub-differential of locally Lipschitz functions. In addition, we propose a type of Wolfe dual problem and explore weak/strong duality relations under L-invexity-infineness hypotheses.