A Feasible SQP Method Using Augmented Lagrangian Function for General Constrained Optimization

An feasible SQP method is proposed to solve general optimization problems with equality and inequality constraints. First, we transform the original problem to an associated simpler problem with only inequality constraints, and the simplified problem is shown to be equivalent to the original problem...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2012 Fifth International Joint Conference on Computational Sciences and Optimization s. 226 - 229
Hlavní autoři: Xiaowei Jiang, Yueting Yang, Yunlong Lu
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 01.06.2012
Témata:
Approximation algorithms
augmented Lagrangian function
Educational institutions
feasible descent algorithm
general constrained optimization
Lagrangian functions
Optimization
SQP
Transforms
Vectors
ISBN:9781467313650, 1467313653
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:An feasible SQP method is proposed to solve general optimization problems with equality and inequality constraints. First, we transform the original problem to an associated simpler problem with only inequality constraints, and the simplified problem is shown to be equivalent to the original problem under the mild condition. Then we use feasible SQP method to solve the latter problem. Here, we use the augmented Lagrangian function to be objective function. At each iteration, multiplier and penalty parameter are updated by the simpler criterion. Numerical experiments are implemented to test the efficiency of the proposed method.
ISBN:9781467313650
1467313653
DOI:10.1109/CSO.2012.57