线性收敛

与变尺度法和共轭梯度法相比，本文方法可保证线性收敛，在PC平台上可基本达到实时效果。

Compared with variable metric method and conjugate gradient method , our algorithm could assure linear convergence , and near real-time performance could be obtained on PC.

本文证明了GR算法线性收敛，提出了一种改进的GR算法。

We prove the linear convergence of GR algorithm , and present an improved one .

前馈网络的一种超线性收敛BP学习算法

Super-Linearly Convergent BP Learning Algorithm for Feedforward Neural Networks

证明了这些修正的算法具有全局收敛性且保持局部二步Q超线性收敛。

It is proved that the improved algorithms possess global and two step Q superlinear convergence properties .

Minimax问题的一个超线性收敛的SQP算法

A superlinearly convergent SQP algorithm for Minimax Problems

证明了这种方法是q-超线性收敛的和大范围收敛的，并给出了与Newton法的数值比较。

The method is shown to be q-superlinear convergent and global convergent , and the numerical comparison with Newton method is given .

约束优化无严格互补的超线性收敛SQP强次可行算法

A Superlinearly Convergent Strongly Sub-Feasible SQP Algorithm for Constrained Optimization Without Strict Complementarity

等式约束优化问题SQP算法的超线性收敛充要条件

Sufficient and Essential Conditions of Super-linear Convergence for SQP Algorithm in Equality Constrained Optimization Proble

在F为一致P-函数情形，证明了算法的全局收敛性、局部超线性收敛性和二次收敛性，对于非退化的线性互补问题仍具有有限步收敛性。

Global convergence , local superlinear convergence , quadratic convergence and finite termination for nondegenerative linear complementarity problems are proved under the assumption that F is a uniform P & function .

一类修正BFGS算法的局部超线性收敛性

Local Q-superlinear Convergence of a Modified BFGS Algorithm

基于Newton迭代法对于求重根具有线性收敛性，给出了加速其收敛的方法以及迭代公式，收敛速度得到了有效的提高。

Considering that the Newton iterative algorithm has first-order convergence in multiple roots , the author advances the method of accelerating its convergence and iteration formula .

在第一章中，对具有一般约束的非线性规划构造出新的具有超线性收敛性的SQP算法。

In the first chapter , introduces a new SQP algorithm for nonlinear optimization with equality and inequality constraint .

本文提出一个求解非线性minimax问题的新算法，该算法具有全局和一步超线性收敛性。

In this paper , a new algorithm for nonlinear minimax problems is presented which has global and one - step superlinear convergence .

并且，算法使用BFGS拟牛顿方法更新矩阵Bk，不需要函数是一致凸的假设条件，我们证明该方法具有全局收敛性和超线性收敛性。

Furthermore , we use BFGS method to update matrix B_k , without the assumption condition of f is uniformly convex , we show that the method is globally and superlinearly convergent .

利用一个修正的BFGS公式，提出了结合线搜索技术的BFGS-信赖域方法，并在一定条件下证明了该方法的全局收敛性和超线性收敛性。

By using a modified BFGS formula , a BFGS-type trust region method with line search technique for unconstrained optimization problems is proposed .

随后，基于EM算法，提出了该模型的新算法EMSFM，证明了该算法的线性收敛法，分析了它的收敛速度。

Accordingly , based on EM algorithm , a new algorithm EM SFM for this model is presented , its linear convergence is proved , and its convergence rate is also analyzed .

本文将集中讨论局部收敛性，特别是证明了在使用DFP或PSB等矩阵校正公式时，修正后的方法在一定的条件下是超线性收敛的。

Particularly , it is proved that if DFP or PSB matrix updating formulae are used , then our method will be convergent superlinearly under some conditions .

本文指出当Jacobian近似满足有界退化性质时，由拟牛顿算法得到的迭代序列最多是线性收敛的。

This paper points out that if a sequence of Jacobian approximation satisfies the bounded deterioration , the sequence generated from the quasi & Newton methods is no better than linear convergence .

POWELL关于拟牛顿法的收敛性的结果推广到超记忆DFP方法，得到了超记忆DFP方法的收敛性定理和超线性收敛速度的结果，并讨论了记忆项的作用。

POWELL are generalized to the supermemory DFP method and its convergence theorem and superlinear rate of convergence are obtained here with some concerned discussions on functions of memory terms .

在合理的条件下，算法具有整体收敛性且两块校正的双边既约Hessian投影法将保持超线性收敛速率。

The global convergence results of the proposed algorithm are proved while maintaining fast local superlinear convergence rate is established by performing a two-piece update of two-side projected reduced Hessian .

D-F-P方法的超线性收敛

Superlinear convergence of the d-f-p method

在第四章，我们考虑约束规划的一个基于SQP方法的信赖域算法，证明了它的全局收敛性和超线性收敛。

Both of them deal with the equality constrained optimization . In Chapter ⅳ, a trust-region algorithm based on SQP method is given , its global convergence and superlinear convergence are proved .

文中对非线性最小二乘问题给出了一种新型求解方法&分块尺度化ABS求解方法，并证明了这类方法的局部超线性收敛性，最后在尺度矩阵取特殊值时给出了数值结果。

As a new type of method for nonlinear least square problems , the block scald ABS meth-ods are presented , and their locally superlinear convergence is proved . Finally the numerical re-sults are given with the special choice of the sealed matrices .

本文研究使用Ishikawa迭代格式求实李普希兹映射的不动点，指出该迭代格式仅具线性收敛率；

In this paper we consider Ishikawa 's iteration scheme to find fixed points of real Lipschitz mappings . We declare that the Ishikawa 's iteration only has linear convergence rate .

提出一个求解LC1无约束优化问题的信赖域算法，在较弱条件下证明了全局收敛性和超线性收敛性。

A trust region method for LC ~ 1 unconstrained problems was presented and under mild conditions we proved its global and superlinear convergene .

该算法全局和超线性收敛并且去掉了传统SQCQP算法全局收敛性分析中的一致正定性假设。

The algorithm is globally and superlinearly convergent , and the uniformly positive definiteness assumption in the global convergence analysis of traditional SQCQP algorithms is removed .

Powell在文献[1]中证明了D-F-P方法是超线性收敛的，Dennis等人在文献[2]也给出了同样的结果，但是证明方法不同。

The D-F-P method was proved to be superlinear convergence in reference [ 1 ] by Powell . Same result was given in reference [ 2 ] by Dennis and More ' but the methods used were different between reference [ 1 ] and [ 2 ] .

并且该方法的收敛速率至少是R-线性收敛的。

The convergence rate of this method is at least R-linearly .

非线性约束最优化一族超线性收敛的可行方法

A Class of Superlinearly Convergent Feasible Methods for Nonlinearly Constrained Optimizations

均衡约束优化具有超线性收敛性算法的研究

Research on the Superlinearly Convergent Algorithms for Optimization with Equilibrium Constraints