On local search in d.c. optimization problems

摘要

First, we consider a d.c. minimization problem with a simple feasible set and develop a special method based on the linearization with respect to the basic nonconvexity. The convergence of the methods is analyzed and compared with published results. Theoretical and practical stopping criteria are proposed. Second, we consider a problem with d.c. constraint and study the properties of special local search method for this problem. Finally, we consider a variant of local search for a general d.c. optimization problem and investigate its convergence.