摘要
将ＡＢＳ算法用于求解线性等式约束的优化问题。给出一个信赖域算法；该算法中用隐式ＬＵ分解算法修正Ｈｅｓｓｅ矩阵，用对称的ＡｂＳ算法求解子问题。证明了由算法生成的序列的任意聚点满足线性等式约束优化问题最优解的必要条件。
The ABS algorithms are utilized to solve a linear equality constrained problem.A trust region algorithm is proposed in which the implicit LU algorithm is applied for modifying Hesse matrices and the subproblems are solved by using the symmetric ABS algorithm, and the authors prove that any cluster point of the sequence generated by the algorithm satisfies the optimal necessary conditions for this problem.
出处
《大连理工大学学报》
CAS
CSCD
北大核心
1994年第3期263-267,共5页
Journal of Dalian University of Technology
基金
国家自然科学基金
关键词
非线性规划
约束问题
最优化算法
nonlinear programming
constrained problem
optimization algorithm/trust region method
ABS algorithm
implicit LU algorithm
