Algorithms in the method of centers with approximation of the set of admissible solutions (Q1405717)

In this paper iterative algorithms for solving mathematical programming problems are considered. In particular verifiable conditions are studied which ensure that the solutions can be attained with the prescribed accuracy within a finite number of iterations. The conditions given are obtained for methods of both interior and exterior centers. In order to obtain the conditions mentioned neighbourhoods and, respectively subsets of the set of feasible points of the initial problem are introduced in the algorithms. It is assumed that the auxiliary unconstrained problems can be solved exactly. The criteria allowing to attain a prescribed accuracy within a finite number of iterations are developed by means of a one-sided (but not two-sided) process of approximation.