An adaptive trust region method and its convergence

In this paper, a new trust region subproblem is proposed. The trust radius in the new subproblem adjusts itself adaptively. As a result, an adaptive trust region method is constructed based on the new trust region subproblem. The local and global convergence results of the adaptive trust region method are proved. Numerical results indicate that the new method is very efficient.     

On generalized hamiltonian systems

1. PreliminaryIt is well known that{1] a 8ymPlectic form is invariant along the trajectory of a Hamilto-nian system. Based on this fundamental property, certain techniques have been developed.The purpose of this paper is to extend such an approach to a wider class of dynamic systeIns,namely, genera1ized Hamiltonian systems. Our purpose is to investigate a class of dynaInicsystems, which possess a certain "geometric structure".Deflnition 1.1[1'2]. Let M be a tIlallifo1d. w E fl'(M) is call…     

极大单调算子的一个新的近似邻近点算法

研究集值映射方程0 T (z)的求解问题, 其中T是极大单调算子.对于给定的x^k及β _k>0, 大部分已有的近似邻近点算法取x^k+1= 满足 x^k +e^k +β_kT(x^k ), ||e^k||≤h_k||x_k- x^k ||, 其中{h^k}为非负可加数列. 新方法中不取 x^k+1 = x^k , 而将新的迭代点取为 x^k+1 = P_Ω [x^k-e^k], 其中Ω 是T的定义域,P_Ω (&#8729;) 表示Ω上的投影算子. 在supk>0h^k < 1这样宽松的条件下给出了收敛性证明.     

A new approximate proximal point algorithm for maximal monotone operator

The problem concerned in this paper is the set-valued equation 0 ∈T(z) where T is a maximal monotone operator. For given xk and βk > 0, some existing approximate proximal point algorithms take x~(k+1) = xk such thatwhere {ηk} is a non-negative summable sequence. Instead of xk+1 = xk , the new iterate of the proposing method is given bywhere Ω is the domain of T and PΩ(·) denotes the projection on Ω. The convergence is proved under a significantly relaxed restriction supk>0 ηk<1.