Nonlinear separation approach to constrained extremum problems (Q1937011): Difference between revisions

The authors consider optimality conditions for cone-constrained optimization problems. Using Hiriart-Urruty's scalarization function they construct several types (nonlinear weak, nonlinear regular weak, nonlinear strong) of separation functions. By applying these separation functions as well as image space analysis, they obtain necessary and sufficient optimality conditions and a corresponding saddle-point condition.