共查询到18条相似文献，搜索用时 390 毫秒
1.
解非线性方程的一类改进型牛顿法 总被引：1，自引：0，他引：1
牛顿迭代法的改进形式主要有算术平均牛顿法(AN)、几何平均牛顿法(GN)、中点牛顿法(MN)、调和平均牛顿法(HN)、α-幂平均牛顿法(PN)等.通过将算术平均牛顿法(AN)与经典牛顿法结合,提出一种新的牛顿型算法,收敛阶可达6阶.与现有算法相比较,该算法具有计算量少、收敛速度快的优点. 相似文献
2.
利用修正牛顿法提出了求解绝对值方程Ax-︱x︱=b的一种算法,对算法的收敛性进行了分析和证明,并将修正牛顿法与牛顿法进行了比较,比较表明修正牛顿法的计算效率更高,最后通过数值实验表明该算法是有效的. 相似文献
3.
拟牛顿法是所有利用一阶导数求解无约束优化问题的方法中最有效的一类计算方法,如何提高实际计算中的运算效率,如何使得对非凸目标函数保持局部超线性收敛的同时具有全局收敛性,是对拟牛顿法进行研究的两个方向.对近年来相关文献的几种修正拟牛顿法进行分析比较,并提出和分析了一个修正BFGS拟牛顿法的收敛性. 相似文献
4.
结合正矢函数,在Fischer-Burmeister函数的框架下给出一种新的二阶锥互补函数.利用该函数设计了一种求解二阶锥互补问题的光滑牛顿法,证明算法具有全局收敛性,并给出了数值实验. 相似文献
5.
讨论了基于Levenberg—Marquardt（LM）算法的BP神经网络及其GDP预测的应用。LM算法利用误差函数二阶导数信息,对高斯一牛顿法的优化,相对传统的负梯度法而言,其收敛速度更快。最后以贺州市GDP为例,就预测的效率和精确度来说,LM—BP网络预测GDP的速度和精度明显优于标准的BP算法网络。 相似文献
6.
对一般目标函数极小化问题,提出一类新的修正阻尼牛顿法.若Hessian矩阵正定且目标函数梯度不为零,则搜索方向取牛顿方向;若Hessian矩阵不正定且非奇异,且目标函数梯度的转置和牛顿方向的数量积大于零时,搜索方向采用负牛顿方向;若Hessian矩阵奇异或者目标函数梯度的转置和牛顿方向的数量积等于零时,搜索方向则采用负梯度方向.因此该算法能保证搜索方向始终为下降方向,并证明对一般的非凸目标函数,该算法全局收敛. 相似文献
7.
基于光滑Fischer-Burmeister函数,给出求解线性对称锥规划的一步光滑牛顿法．该算法在每一步迭代只需求解一个线性方程组,并进行一次线性搜索．不必满足严格互补,算法具有全局收敛性． 相似文献
8.
通过对现有的二维盲图像恢复算法的探讨,提出了两种基于L1双正则化的二维盲图像恢复算法.一种是最小化L2-L1代价函数,为了实现边缘保持和噪声抑制;另一种是通过最小化L1-L1代价函数来处理非高斯噪声的情况.所提的算法是一种广义的梯度算法,它通过引入绝对值函数的弱导数来处理不可微的情况.实验结果表明,与NAS-RIF算法和DR算法相比,所提出的两种二维算法能够更快速地获得好的图像估计. 相似文献
9.
10.
11.
蔡春 《唐山师范学院学报》2004,26(2):20-25,60
提出了牛顿方法与预优广义共轭梯度方法相结合的方法(简称为Newton-PGCG)求解库恩-塔克方程组。给出Newton-PGCG算法中预优广义共轭梯度法中预优阵的选取问题及其步数参数的确定原则，并在保证牛顿方法恰二阶收敛的条件下，证明了此算法也具有恰二阶收敛的优点。 相似文献
12.
* In the separation operations of process industry,we often encounter some systems in which the relativevolatilities between components are close to one,or azeotropes are generated, and it is unreasonable ineconomy or unfeasible in technology that the conven-tional distillation is used to separate these systems. Ifthe azeotropic distillation is utilized for separatingthese mixtures, the knowledge about the compositionand boiling temperature of the azeotropes must be wellknown. The accurate pre… 相似文献
13.
Based on the generalized Fischer-Burmeister function, Chen et al in 2008 put forward a regularization semismooth Newton method for solving the nonlinear complementarity problem with a p0 -function. In this paper, we investigate the above algorithm with the monotone line search replaced by a non-monotone line search. It is shown that the non-monotone algorithm is well-defined, and is globally and locally superlinearly convergent under standard as- sumptions. 相似文献
14.
基于CHKS光滑函数,将非线性互补问题转化为非线性光滑方程组,再构造光滑算子,将非线性光滑方程组转化为优化问题,且构造了一个新的牛顿算法,该算法引入了非单调线搜索,并在一定条件下证明了它的全局收敛性,及在非奇异条件而非严格互补条件条件下,证明了它的局部二次收敛性。最后给出数值实验结果。 相似文献
15.
By using the theory of Euclidean Jordan algebras,based on a new class of smoothing functions,the QiSun-Zhou's smoothing Newton algorithm is extended to solve linear programming over symmetric cones (SCLP).The algorithm is globally convergent under suitable assumptions. 相似文献
16.
By using a smoothing function,the P nonlinear complementarity problem(P NCP)can be reformulated as a parameterized smooth equation.A Newton method is proposed to solve this equation.The iteration sequence generated by the proposed algorithm is bounded and this algorithm is proved to be globally convergent under an assumption that the P NCP has a nonempty solution set.This assumption is weaker than the ones used in most existing smoothing algorithms.In particular,the solution obtained by the proposed algorithm is shown to be a maximally complementary solution of the P NCP without any additional assumption. 相似文献
17.
The algorithm for the approximate solution of a geodesic connecting two given points on a simple surface is discussed in this paper. It arises from practical demands of the filament winding technique. Geodesic is the shortest path connecting two given points on a surface and it can also be regarded as the extremal curve of the arc length functional. The nonlinear equation system of the geodesic on some discrete points by means of the direct variation method is explored. By employing Newton’s iterative method, this nonlinear system is transformed into a linear one. And the approximate solution to the geodesic is obtained by solving the resultant linear system. This paper also proves that the iteration is convergent under certain circumstance. Moreover, the result is illustrated with three examples and an appropriate comparison between the analytical solution and the approximate solution to the geodesic is described on the cone surface. 相似文献
18.
在经典牛顿法的基础上,给出了求解非线性方程组的非精确牛顿法。在一定的条件下,证明了该算法的超线性收敛性,并且这个收敛性是二阶的。 相似文献