结合非单调信赖域方法,和非单调线搜索技术,提出了一种新的无约束优化算法.信赖域方法的每一步采用线搜索,使得迭代每一步都充分下降加快了迭代速度.在一定条件下,证明了算法具有全局收敛性和局部超线性.收敛速度.数值试验表明算法是十分有效的.
In this paper, we combine non-monotonic trust region method and non-monotonicline search technique to propose a new algorithm for unconstrained optimization. Each stepof trust region method employs line search, make each step of iteration fully decrease, thenaccelerate the iterative speed. Global convergence and local superlinear convergence rate isproved under certain conditions. Preliminary numerical results show the algorithm is effective.