摘要
提出了一个等式约束下凸二次规划问题的带强Wolfe线搜索的信赖域算法.该算法利用增广Lagrange函数将约束问题转化为无约束问题,在传统信赖域算法的基础上结合线搜索技术,当试探步失败时不用重解信赖域子问题,减少了计算量.在适当的条件下,证明了此算法的全局收敛性.
A new trust region algorithm with strong Wolfe line search for convex quadratic programming under equality constraints is proposed.This method first uses augmented Lagrange function to transform this restraint question into the non-constraint question,and on the basis of traditional trust region method it takes line search instead of resolving the subproblem when the trail step is not successful.This may allow a considerable computational saving.Global convergence is proved under certain conditions.
出处
《郑州轻工业学院学报(自然科学版)》
CAS
2011年第4期117-120,共4页
Journal of Zhengzhou University of Light Industry:Natural Science