An aggregate subgradient method for nonsmooth and nonconvex minimization (Q1082780)

This paper presents an algorithm for minimizing a locally Lipschitz continuous function which is not necessarily convex. The given convergence results are not so strong especially the global convergence result which is established under the hypothesis of convexity of the function to be minimized. Concerning the proofs the reader is often referred to the bibliography. The paper ends with an appendix giving two numerical tests. The considered functions have only two variables. The author does not compare his algorithm with existing ones. In fact the nonsmooth optimization is a hard field of investigation. This explains the weakness of the obtained results.