CAJ | 학술논문

本文对线性约束优化问题提出了一个新的广义梯度投影法,该算法采用了非精确线性搜索,并在每次迭代运算中结合了广义投影矩阵和变尺度方法的思想确定其搜索方向.在通常的假设条件下,证明了该算法的整体收敛性和超线性收敛速度.
본문대선성약속우화문제제출료일개신적엄의제도투영법,해산법채용료비정학선성수색,병재매차질대운산중결합료엄의투영구진화변척도방법적사상학정기수색방향.재통상적가설조건하,증명료해산법적정체수렴성화초선성수렴속도.
In this paper, a new generalized gradient projection method with inexact line search is proposed for the nonlinear optimization problem with linear constraints. The main idea is to use the generalized projection matrix combined with the variable metric method to determine the search direction at each iteration. Under some suitable assumption it is proved that our method is of global convergence properties and superlinear convergence rate.