ДЕЯКІ ПІДХОДИ ДО РЕГУЛЯРИЗАЦІЇ НЕЛІНІЙНИХ ЗАДАЧ ОПТИМІЗАЦІЇ

Methods for transformation of convex optimization problems with constraints into equivalent problems with better computational properties are considered. Main attention is paid to conical approximations and conical extensions of the objective functions from the feasible set to the full space of vari...

Full description

Saved in:
Bibliographic Details
Date:2025
Main Authors: Laptin, Yu.P., Bardadym, T.A.
Format: Article
Language:English
Published: V.M. Glushkov Institute of Cybernetics of NAS of Ukraine 2025
Online Access:https://jais.net.ua/index.php/files/article/view/568
Tags: Add Tag
No Tags, Be the first to tag this record!
Journal Title:Problems of Control and Informatics

Institution

Problems of Control and Informatics
Description
Summary:Methods for transformation of convex optimization problems with constraints into equivalent problems with better computational properties are considered. Main attention is paid to conical approximations and conical extensions of the objective functions from the feasible set to the full space of variables. As a result we get an unconstrainerd (regularized) convex programming problem, whose solution coincides with the solution of the initial one. Special importance these approaches have in the case when the objective function is not defined outside of the feasible set. Effective procedures for computation of auxiliary functions are proposed, peculiarities of software realizations are considered, results of computational experiments are reported.