Zobraziť v EDS

A superlinearly convergent feasible descent algorithm for nonlinear optimization

Uložené v:
Podrobná bibliografia
Názov: A superlinearly convergent feasible descent algorithm for nonlinear optimization
Autori: Jian, Jinbao
Informácie o vydavateľovi: Wuhan University, Wuchang
Predmety: superlinearly convergent feasible descent algorithm, Applications of renewal theory (reliability, demand theory, etc.), Nonlinear programming
Popis: The author considers the following problem: \((p)\) \(\min \{f_0 (x)|x \in R\}\), where \(R = \{x \in E^n |f_i (x) \leq 0\), \(j \in L_1\); \(f_j (x) = 0\), \(j \in L_2\}\), and discusses optimization with nonlinear equality and inequality constraints. The author presents a new algorithm possessing the following properties: (1) The algorithm is a feasible descent method for the expansive problems and the parameter adjusts automatically only for finite time; (2) only one quadratic program needs to be solved at each iteration; (3) It superlinearly converges to the solution of the original problem under some suitable assumption.
Druh dokumentu: Article
Popis súboru: application/xml
Prístupová URL adresa: https://zbmath.org/883157
Prístupové číslo: edsair.c2b0b933574d..0c8c93bc143425ade3a62a914d217c67
Databáza: OpenAIRE
Popis
Abstrakt:The author considers the following problem: \((p)\) \(\min \{f_0 (x)|x \in R\}\), where \(R = \{x \in E^n |f_i (x) \leq 0\), \(j \in L_1\); \(f_j (x) = 0\), \(j \in L_2\}\), and discusses optimization with nonlinear equality and inequality constraints. The author presents a new algorithm possessing the following properties: (1) The algorithm is a feasible descent method for the expansive problems and the parameter adjusts automatically only for finite time; (2) only one quadratic program needs to be solved at each iteration; (3) It superlinearly converges to the solution of the original problem under some suitable assumption.