共轭梯度法教程.docxVIP

PAGE  《非线性规划》课程设计 题 目： 共轭梯度法及其数值实现 院 系： 数理科学与工程学院应用数学系 专 业： 数学与应用数学 姓名学号： *** ********* *** ********* *** ********* *** ********* 指导教师： 张世涛 日 期： 2015 年 7 月 5 日 摘要 最优化是一门应用性很强的学科，近年来，随着计算机的发展以及实际问题的需要，大规模优化问题越来越受到重视。共轭梯度法是最优化中最常见的方法之一，他具有算法简单、存储需求少、有较快的收敛速度和二次终止性且易于实现等优点，十分适合于大规模优化问题。 非线性共轭梯度法已有五十多年的历史，最早由计算数学家Hestenes和几何学家Stiefel为求解线性方程组Ax=b，而独立提出的。较著名的有FR方法、PRP方法、HS方法和LS方法等。非线性最优化的共轭梯度算法的收敛性分析，也就是讨论各种共轭梯度算法在不同搜索下的收敛性质。本文主要研究求解无约束优化问题的非线性共轭梯度法，并用Matlab软件对其数值实现。 关键词：无约束规划；非线性共轭梯度法；迭代；最优解；数值实现 Abstract Optimization is strong discipline applied. In recent years, with the development of computer and practical issues,large-scale optimization problems are given more and more attention. Conjugate gradient method is one of the most commonly used methods in optimization. It is simply, storage needs less, easy to practice with faster convergence speed and quadratic termination. It is suitable for large-scale optimization problem. The conjugate gradient method have been more than 50 years of history. The pioneers were mathematician Hestenes and geometrician Stiefel. They independently proposed this method for solving system of linear equations Ax=b，. Well-known conjugate gradient method is FR method, PRP method, HS method, LS method and so on. The convergence analysis of the conjugate gradient algorithm for nonlinear optimization is also the convergence of various conjugate gradient algorithms under different search conditions. Global convergence and numerical result of nonlinear conjugate gradient method of unconstrained optimization is investigated in this paper. Besides, we use Matlab to get its numerical solution. Keywords: Unconstrained programming; Nonlinear conjugate gradient method; Iteration; Optimal solution; Numerical implementation非线性规划课程设计  PAGE \* MERGEFORMAT 10 目录 第一章 引言………………………………………………………………………2 1.1无约束优化问题概述………………………………………………………2 1.2 共轭方向……………………………………………………………………2 1.3 共轭方向法…………………