Directional Kuhn--Tucker condition and duality for quasidifferentiable programs (Q1422313)

The focus in the paper is on necessary and sufficient optimality conditions for quasidifferentiable optimization problems with inequality constraints. Being obvious for quasidifferentiable programs the optimality conditions are formulated in terms of the directional derivative of the (regular) Lagrange function of the problem. In the necessary conditions it is shown that direction depending Lagrange multipliers exist such that this directional derivative is non-negative for all directions. If all problem functions are \(\eta\)-invex, the necessary condition is also sufficient for global optimality. Examples show that the Lagrange multipliers really depend on the direction for non-convex problems.