Evgenij G. Golstejn; Thomas Unger : Generalized Level Method with Approximated Data

Author(s) :

Evgenij G. Golstejn; Thomas Unger

Title :

Generalized Level Method with Approximated Data

Preprint series

Technische Universität Chemnitz, Fakultät für Mathematik (Germany). Preprint 2000-5, 2000

Mathematics Subject Classification :

90C25 [ Convex programming ]

65K05 [ Mathematical programming (numerical methods) ]

Abstract :

We describe a generalization of the level method for solving a nondifferentiable program where the problem data (feasible set, functionvalues, and subgradients) are not known explicitely and may be approximated using an oracle only up to an appropriately chosen accuracy. We show that it produces an approximately feasible and optimal solution after a finite number of oracle calls.

Keywords :

nondifferentiable optimization, convex optimization, level method, cutting plane method, inexact data

Language :

english

Publication time :

5/2000