论文部分内容阅读
提出一个带线性搜索的非单调自适应信赖域算法。该算法将非单调 wolfe线搜索和自适应信赖域方法相结合，同时采用新的BFGS校正公式，保持了信赖域子问题海森矩阵的正定性。在适当条件下，分析了该算法的全局收敛性，并通过数值实验说明了该算法的可行性。“,”In this paper,a new self-adaptive trust region method with nonmonotone line search is proposed.The presented method combines non-monotone wolfe line search with self-adaptive trust region method and adopts the modified quasi-Newton equation to ensure the positive-defi-niteness for Hessian matrix of trust region subproblem.By analyzing the global convergence un-der appropriate conditions,it aims to prove the feasibility of the new method through numerical experiments.