作者: Jie Shen , Li-Ping Pang , Dan Li
DOI: 10.1155/2013/697474
关键词:
摘要: An implementable algorithm for solving a nonsmooth convex optimization problem is proposed by combining Moreau-Yosida regularization and bundle quasi-Newton ideas. In contrast with methods of Mifflin et al. (1998), we only assume that the values objective function its subgradients are evaluated approximately, which makes method easier to implement. Under some reasonable assumptions, shown have Q-superlinear rate of convergence.