A robustification approach in unconstrained quadratic optimization
成果类型：
Article
署名作者：
Bernauer, Martin K.; Griesse, Roland
署名单位：
Technische Universitat Chemnitz
刊物名称：
MATHEMATICAL PROGRAMMING
ISSN/ISSBN：
0025-5610
DOI：
10.1007/s10107-009-0302-9
发表日期：
2011
页码：
231-252
关键词：
摘要：
Unconstrained convex quadratic optimization problems subject to parameter perturbations are considered. A robustification approach is proposed and analyzed which reduces the sensitivity of the optimal function value with respect to the parameter. Since reducing the sensitivity and maintaining a small objective value are competing goals, strategies for balancing these two objectives are discussed. Numerical examples illustrate the approach.