Spectral Projected Gradient Method with Inexact Restoration for Minimization with Nonconvex Constraints

This work takes advantage of the spectral projected gradient direction within the inexact restoration framework to address nonlinear optimization problems with nonconvex constraints. The proposed strategy includes a convenient handling of the constraints, together with nonmonotonic features to speed up convergence. The numerical performance is assessed by experiments with hard-spheres problems, pointing out that the inexact restoration framework provides the adequate environment for the extension of the spectral projected gradient method for general nonlinearly constrained optimization.