摘要
对于无约束最优化问题minf(x),x∈Rn,提出了一种广义拟牛顿算法,并且讨论了广义拟牛顿算法对一般目标函数的全局收敛性,以及当f(x)满足Lipschitz连续的条件下,证明了相应的超线性收敛定理。
A class of algorithms which are update Quasi-Newton methods for unconstrained optimization are as follows： min f（ x ）, x ∈ R^n. The proof of the global and superlinearly convergence of the generalized-quasi-Newton methords is proposed. Three main theorems are given.
出处
《北京联合大学学报》
CAS
2007年第4期69-70,76,共3页
Journal of Beijing Union University
关键词
广义拟牛顿算法
全局收敛性
超线性收敛性
generalized-quasi-Newton algorithms
superlinearly convergence
uncontrained optimization
