Выходит 12 номеров в год
ISSN Печать: 1064-2315
ISSN Онлайн: 2163-9337
Indexed in
Some Approaches to Regularization of Nonlinear Optimization Problems
Краткое описание
We consider ways of transformation of convex optimization problems with constraints to equivalent problems with better computational properties. Sufficient attention is paid to use of conic approximations and conic prolongation of objective functions from admissible domain of optimization problem on the whole space of variables. The problem of convex programming without constraints (regularized problem), which solution coincides with solution of the initial problem, is the result of usage of the suggested approach. The considered approaches have special significance when the objective function is not defined outside the admissible domain. We suggested efficient procedures for calculation of auxiliary functions, consider peculiarities of software support of algorithms, results of computational experiments.
-
Laptin Yu. P., Exact Penalty Functions and Convex Extensions of Functions in Schemes of Decomposition in Variables*, Cybernetics and Systems Analysis, 52, 1, 2016. Crossref
-
Laptin Yu. P., Berezovskyi O. A., Using Conical Regularization in Calculating Lagrangian Estimates in Quadratic Optimization Problems, Cybernetics and Systems Analysis, 53, 5, 2017. Crossref