Copositivity meets D. C. optimization

Abstract

Based on a work by M. Dur and J.-B. Hiriart-Urruty[3], we consider the problem of whether a symmetric matrix is copositive formulated as a difference of convex functions problem. The convex nondifferentiable function in this d.c. decomposition being proximable, we then apply a proximal-gradient method to approximate the related stationary points. Whereas, in [3], the DCA algorithm was used.