结合CF和PCG搜索的拟牛顿优化算法

Quasi-Newton optimization algorithm combined with CF and PCG search

针对拟牛顿优化算法求解非线性方程组和无约束优化问题时,需要进行大量的迭代计算的问题.文中提出了一种结合CF和PCG搜索的拟牛顿优化算法,该算法结合CF和PCG搜索的步长因子来得到一种有效的牛顿搜索算法.在强Wolfe准则下的全局收敛性和数值分析结果表明,文中所提出的算法能加快拟牛顿优化算法的求解速度并能得到更高的精度.

When quasi-Newton optimization algorithm is used to solve the nonlinear equations and unconstrained optimiza- tion problems,a large number of iterative computation are needed. Therefore,a quasi-Newton optimization algorithm combined with CF and PCG search is proposed. This algorithm combines the step factor of CF and PCG search to obtain an effective Newton search algorithm. The global convergence and numerical analysis results obtained by means of the strong Wolfe criterion show that the proposed algorithm can accelerate the solution of the quasi-Newton optimization algorithm and obtain high precision.