An Adaptive Proximal Bundle Method with Inexact Oracles for a Class of Nonconvex and Nonsmooth Composite Optimization

In this paper, an adaptive proximal bundle method is proposed for a class of nonconvex and nonsmooth composite problems with inexact information. The are the sum finite convex function information function. For function, we design convexification technique ensure linearization errors its augment to be nonnegative. Then, regarded as approximate primal problem. adopt disaggregate strategy regard cutting plane models model give method. Meanwhile, information, utilize noise management update parameter reduce influence can obtain solution. Two polynomial functions six DC referred in numerical experiment. preliminary results show that our algorithm effective reliable.