一个无约束信赖域梯度路径法的收敛性证明与数值试验

A proof on Trust Region Problems Arising from Unconstrained Optimization and Data Experiment

给出文献[1]中梯度路径法的收敛分析与证明，分析表明方法全局收敛，且在极限点处当海色矩阵正定时，方法满足二次收敛的特性；最后给出算法的数值试验结果．

The paper presents, the global convergence of the analysis and proof in gredient path, The Analysis illustrates the global convergence and atthe stationary point when Hessian matrix is positive in the optional solution,The Mathods satisfies a second convergence property,eventually gives the dade expenmental results of caculating methods'