Lagrangian Approach to Quasiconvex Programing

For a mathematical programming problem, we consider a Lagrangian approach inspired by quasiconvex duality, but as close as possible to the usual convex Lagrangian. We focus our attention on the set of multipliers and we look for their interpretation as generalized derivatives of the performance function associated with a simple perturbation of the given problem. We do not use quasiconvex dualities, but simple direct arguments.