Separation approach for augmented lagrangians in Constrained nonconvex optimization (Q963656)

The authors consider a finite-dimensional inequality-constrained optimization problem. Main results of the image space analysis are recalled (see \textit{F. Gianessi} [Constraint optimization and image space analysis, Springer Verlag, Berlin (2005)]). By means of this theory, local and global saddle point conditions for a general class of augmented Lagrangian functions under suitable conditions are proved. The existence of the global saddle-point does not require the compactness of the feasible set.