一族共轭梯度法的全局收敛性

提出了求解无约束优化问题的一族共轭梯度法，这族方法包古Fkcher提出的共轭下降法．文中证明了一种非精确线性搜索条件能够保证这族方法的下降性和全局收敛性，其收敛结果与Dai和Yuan 1996年给出的关于共轭下降法的相一致．     

结合广义Armijo步长搜索的一类新的共轭度算法及其收敛特征

对求解无约束规划的共轭梯度算法中共轭梯度方向中的参数给了一个假设条件。从而确定它的一个取值范围，使其在此范围内取值均能保证共轭梯度方向是目标函数的充分下降方向。提出了一类新的共轭梯度算法，在去掉迭代点列有界和广义Armijo步长搜索下讨论了算法的全局收敛性。同时给出了具有好的收敛性质和较快收敛速度的FR，PR，HS共轭梯度法的修正形式。数值例子表明新算法比Armijo搜索下的FR，PR，HS共轭梯算法更稳定更有效。算法需要较小的存储。特别适于求解大规模无约束最优化问题。     

一个充分下降的修正 PRP 型谱共轭梯度法

谱共轭梯度法是共轭梯度法的一种重要延拓,可以通过共轭参数和谱参数二维度调整,使得所设计算法的搜索方向满足某一预设条件,比如充分下降条件或共轭条件等。谱参数和共轭参数的设计是谱共轭梯度法的两大核心工作,决定方法的收敛性和数值效果。基于 PRP 方法,构造了一个修正的 PRP 型共轭参数,该共轭参数不仅保持了 PRP 公式的结构和性能,而且具有 FR 方法的收敛性质。利用充分下降条件取定一个谱参数,与修正的 PRP 型共轭参数结合,建立一个新的谱共轭梯度算法。该算法不依赖于任何线搜索就可以满足充分下降条件。常规假设条件下,采用强 Wolfe 线搜索准则产生步长,证明了新算法的全局敛性。通过 100 个算例对该算法进行数值测试并与其他五个算法进行比较,同时采用性能图对数值结果进行直观展示,结果表明该算法是有效的。     

Armijo型线搜索下一种共轭梯度法的收敛性

对无约束非线性规划问题,本文分别在两种不同的Armijo型线搜索下证明了Liu-Storey共轭梯度法的所有搜索方向都是充分下降的,并进一步证明了该算法是全局强收敛的。对另一种放松了函数值下降条件可以获得更大步长的Armijo型线搜索,本文还证明了该算法是全局强收敛的。     

一类修正Zhang H.C.非单调共轭梯度算法

为有效求解大规模无约束优化问题,本文基于RMFI共轭梯度法,结合Zhang H.C.非单调线搜索步长规则,提出了一类新的共轭梯度算法.在适当的条件下,证明了新算法的全局收敛性.数值算例表明,新算法比Zhang H.C.非单调规则下的标准RMFI方法收敛速度更快,更有效.同时,本文进一步研究了Zhang H.C.非单调线搜索步长规则的一个基于强迫函数的拓展模型,并从理论上证明了基于此拓展模型的新算法的全局收敛性.     

结合广义Armijo步长搜索的一类记忆梯度算法及其收敛特征

本文对求解无约束规划的超记忆梯度算法中线搜索方向中的参数,给了一个假设条件,从而确定了它的一个新的取值范围,保证了搜索方向是目标函数的充分下降方向,由此提出了一类新的记忆梯度算法.并在去掉迭代点列有界和广义Armijo步长搜索下,讨论了算法的全局收敛性,且给出了结合形如共轭梯度法FR,PR,HS的记忆梯度法的修正形式,数值实验表明,新算法比Armijo线搜索下的FR,PR,HS共轭梯度法和超记忆梯度法更稳定、更有效.     

在Armijo型线搜索下共轭下降法的收敛性

我们提出了两种Armijo型线搜索,进而证明了这两种Armijo型线搜索可保证共轭下降法的下降搜索方向的充分下降性。并在这两种Armijo型线搜索下得到共轭下降法的收敛性结果。     

广义共轭梯度法及收敛条件

本文对广义共轭梯度法给出了两个简单的收敛条件,这两个条件比[1]的条件弱,并据此给出了两族模型算法和两个特定算法,最后证明了它们的收敛性。     

一类Armijo搜索下的混合HS-PRP共轭梯度法?

为有效求解大规模无约束优化问题,本文基于HS方法和PRP方法,提出了一类新的混合共轭梯度法。该方法在每步迭代中都不依赖于函数的凸性和搜索条件而自行产生充分下降方向。在精确搜索下,本文算法将还原为标准的PRP方法。在适当的条件下,获证了该法在Armijo搜索下,即使求解非凸函数极小化的问题,算法也具有全局收敛性。同时,数值实验表明本文算法可以有效求解优化测试问题。     

非线性约束条件下一类采取曲线搜索的可行方向法

本文推广了无约束最优化方法中采用曲线搜索这一结果,从而给出了一类非线性约束条件下来用曲线搜索的可行方向法。并且在一定的条件下,我们证明了此算法类是全局收敛的. 关键词:非线性约束;;下降可行方向对;;曲线搜索;;可行方向法;;收敛性。     

求解非线性等式约束优化问题的共轭梯度投影算法

利用投影矩阵,对求解无约束规划的共轭梯度算法中的参数βk给一限制条件确定βk的取值范围,以保证得到目标函数的共轭梯度投影下降方向,建立了求解非线性等式约束优化问题的共轭梯度投影算法,并证明了算法的收敛性。数值例子表明算法是有效的。     

The optimum design of electrode and insulator contours by nonlinearprogramming using the surface charge simulation method

A new method is presented for optimizing electrode and insulator contours. The contours are modified by using the iteration methods of nonlinear programming until the desired electric field distribution is obtained. The Gauss-Newton, quasi-Newton, conjugate gradient, or steepest descent method is used for the iteration. The electric-field distributions are computed by means of the surface charge simulation method. It is shown that the Gauss-Newton method gives very fast convergence     

Magnetic resonance imaging gradient coil design by combiningoptimization techniques with the finite element method

In this paper, the optimization techniques of complex method, steepest descent, and conjugate gradient are investigated in terms of their convergence behaviors. The conjugate gradient method is then combined with finite element analysis techniques to develop a magnetic resonance imaging (MRI) G_z gradient coil design strategy which maximizes the field linearity within a specified region of interest. It is found that conjugate gradient optimization in conjunction with the finite element method is a powerful and flexible coil design approach with the potential to incorporate complex coil geometries, inhomogeneous media, and transient current excitation     

共轭梯度法的全局收敛性

探讨了在强Wolfe搜索规则下，与βk^PR相关的算法的收敛性，在不需要假设目标函数为凸的情况下，证明了充分下降及算法的全局收敛性。     

The solution of sparse linear equations by the conjugate gradient method

The convergence properties of the conjugate gradient method are discussed in relation to relaxation methods and Chebyshev accelerated Jacobi iteration when applied to the solution of large sets of linear equations which have a sparse, symmetric and positive definite coefficient matrix. The conclusion is reached that its convergence rate is unlikely to be much worse than these methods, and may be considerably better. The conjugate gradient method may either be applied to the basic unscaled or scaled equations or alternatively to various transformed equations. Preconditioning, block elimination and partial elimination methods of transforming equations are considered, and some comparative tests given for six problems.     

Study on semi‐conjugate direction methods for non‐symmetric systems

Some theoretical problems and implementation problems are studied here for the semi‐conjugate direction method established by Yuan, Golub, Plemmons and Cecilio (2002). The existence of semi‐conjugate directions is proved for almost all matrices except skew‐symmetric matrices. A new technique is proposed to overcome the breakdown problem appeared in the semi‐conjugate direction method. In the implementation of the semi‐conjugate direction method, the generation of the semi‐conjugate direction is very important and necessary, but very expensive. The technique of limited‐memory is introduced to economize the cost of the generation of the semi‐conjugate direction in the Yuan–Golub– Plemmons–Cecilio algorithm. Finally, some numerical experiments are given to confirm our theoretical results. Our results illustrate that the semi‐conjugate direction method is very nice alternative for solving non‐symmetric systems, and the limited‐memory left conjugate direction method is a good improvement of the left conjugate direction method. Copyright © 2004 John Wiley & Sons, Ltd.     

Least-squares finite element method and preconditioned conjugate gradient solution

A least-squares variational procedure for first-order systems of differential equations and an approximate formulation based on finite elements are developed. Error estimates, a condition number bound and analysis of weighting factors are given. Steepest descent and conjugate gradient solution procedures are examined, and an appropriate preconditioner constructed which is demonstrated to yield rapid convergence and to be insensitive to problem size. Numerical studies of rates of convergence for a test problem are given.