类Broyden族非拟牛顿算法对一般目标函数的全局收敛性

Global Convergence of Similar Quasi-Newton Method for Nonconvex Unconstrained Optimization Problems

应用双参数的类Broyden族校正公式,为研究求解无约束最优化问题的拟牛顿类算法对一般目标函数的收敛性这个开问题提供了一种新的方法．

In this paper, we propose a similar quasi-Newton update with two parameters for nonconvex unconstrained optimization problems, and prove that the method with Goldstein line search converges globally.