Numerical methods for nonsmooth DC Optimization and their applications
Adil Bagirov
Federation University Australia, Ballarat, Victoria, Australia
In this talk, we present numerical methods for solving nonsmooth difference of convex (DC)
optimization problems. We discuss various generalized subdifferentials of nonsmooth DC functions
and formulate optimality conditions using them. We consider numerical methods which use both
explicit and implicit DC representations. These methods are extensions of subgradient and bundle
methods. Results of numerical experiments using nonsmooth DC optimization academic test problems
are reported. The application of these methods for solving piecewise linear regression problems is
demonstrated.