Xia Dong-jing
Abstract
The filter method is a class of effective method for nonlinearly constrained optimization.The penalty function is not employed.A new rule is given to accept a trial point in the filter method.Based on the rule,a new algorithm is proposed,which is combined with SQP algorithm.Under some basic assumptions,the algorithm is globally convergent.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
The filter method is a class of effective method for nonlinearly constrained optimization.The penalty function is not employed.A new rule is given to accept a trial point in the filter method.Based on the rule,a new algorithm is proposed,which is combined with SQP algorithm.Under some basic assumptions,the algorithm is globally convergent.
Key concepts: Sequential quadratic programming, Filter (signal processing), Algorithm, Convergence (economics), Penalty method, Mathematical optimization, Mathematics, Adaptive filter