Newton Method for Optimization Problem with Equal Constrains Can be Improved
Cai Chun
Abstract
Cai Chun
Abstract
The Newton-PGCG algorithm is proposed to solve the Kuhn-Tucker equations. It is such a new algorithm that every Newton step is followed by some Preconditioned Generalized Conjugate Gradient steps. The steps parameter is given in this paper. Newton-PGCG algorithm is proved quadratically convergent exactly under the same conditions ensuring the Newton algorithm quadratically convergent exactly.
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 Newton-PGCG algorithm is proposed to solve the Kuhn-Tucker equations. It is such a new algorithm that every Newton step is followed by some Preconditioned Generalized Conjugate Gradient steps. The steps parameter is given in this paper. Newton-PGCG algorithm is proved quadratically convergent exactly under the same conditions ensuring the Newton algorithm quadratically convergent exactly.
Key concepts: Quadratic growth, Newton's method, Mathematics, Newton's method in optimization, Conjugate gradient method, Convergence (economics), Newton fractal, Applied mathematics