Quasi-monotone subgradient methods for nonsmooth convex minimization. (English) Zbl 1330.90078
The authors present new subgradient methods for solving nonsmooth convex minimization problems in a finite dimensional setting. It is shown that, by using special multiple averaging schemes, these methods generate convergent minimizing sequences. Applications to primal-dual problems and numerical experiments are provided.
Reviewer: Nicolae Popovici (Cluj-Napoca)
MSC:
| 90C25 | Convex programming |
| 90C47 | Minimax problems in mathematical programming |
| 68Q25 | Analysis of algorithms and problem complexity |
Keywords:
convex optimization; nonsmooth optimization; subgradient methods; multiple averaging; rate of convergence; primal-dual methodsReferences:
| [1] | Shor, N.Z.: Minimization Methods for Non-differentiable Functions. Springer, Berlin (1985) · Zbl 0561.90058 · doi:10.1007/978-3-642-82118-9 |
| [2] | Polyak, B.T.: Introduction to Optimization. Software Inc., New York (1987) · Zbl 0708.90083 |
| [3] | Nemirovsky, A.S., Yudin, D.B.: Problem Complexity and Method Efficiency in Optimization. Wiley, New York (1983) · Zbl 0501.90062 |
| [4] | Nesterov, Yu.: Primal-dual subgradient methods for convex problems. Math. Program. 120, 261-283 (2009) · Zbl 1191.90038 · doi:10.1007/s10107-007-0149-x |
| [5] | Beck, A., Teboulle, M.: Mirror descent and nonlinear projected subgradient methods for convex optimization. Oper. Res. Lett. 31, 167-175 (2003) · Zbl 1046.90057 · doi:10.1016/S0167-6377(02)00231-6 |
| [6] | Nesterov, Yu.: Introductory Lectures on Convex Optimization. Kluwer, Boston (2004) · Zbl 1086.90045 · doi:10.1007/978-1-4419-8853-9 |
| [7] | Rockafellar, R.T.: Convex Analisys. Princeton University Press, Princeton (1970) · Zbl 0193.18401 |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
