A new nonlinear ABS-type algorithm and its efficiency analysis*
作者:
N. Deng,
Z. Chen,
期刊:
Optimization Methods and Software
(Taylor Available online 1998)
卷期:
Volume 10,
issue 1
页码: 71-85
ISSN:1055-6788
年代: 1998
DOI:10.1080/10556789808805702
出版商: Gordon and Breach Science Publishers
关键词: Nonlinear system of equations;ABS-type algorithm;R-order convergence;efficiency
数据来源: Taylor
摘要:
As a continuation work following [4] and [5], a new ABS-type algorithm for a nonlinear system of equations is proposed. A major iteration of this algorithm requiresncomponent evaluations and only one gradient evaluation. We prove that the algorithm is superlinearly convergent withR-order at least τn, where τnis the unique positive root ofτn−τn−1−1=0. It is shown that the new algorithm is usually more efficient than the methods of Newton, Brown and Brent, and the ABS-type algorithms in [1], [4] and [5], in the sense of some standard efficiency measure.
点击下载:
PDF (447KB)
返 回