方程近似解的计算复杂性

来源 :浙江大学 | 被引量 : 0次 | 上传用户:tklyzh1
下载到本地 , 更方便阅读
声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架
论文部分内容阅读
The problem on optimization of approximate solution of operator equations is mainly to determine the exact orders of error and complexity of algorithms,and construct the optimal algorithm realizing the orders.Complexity of operations is also called the cost of operation.Therefore,the problem of the ε-complexity of approximate solution of operator equations,roughing speaking,is the minimal cost among all algorithms which solve the problem with error at most ε,and it has widely practical backgrounds.Generally speaking,the optimization of a problem can be done in various setting,such as worst case setting,average case setting,probability case setting.The worst case setting is in common use,but the others are also of interest.To now,as far as the optimization of approximate solution of operator equations is concerned,many works have been done in the worst case setting.In the average case setting,however,the result about this problem has hardly ever been seen.In Chapter 1 we have considered the problem of optimization of approximate solution of integral equations of several variables in worst case.We determine the exact error order and construct the optimal error algorithm.In Chapter 2 The problem of ε-complexity of integral equations have been considered in worst case.We determine the exact order of the complexity and construct the optimal algorithm realizing the order.In Chapter 3 we have considered the problem of the optimization of approximate solution of operator equations in the average case and its application.To solve above mentioned problems we have used the classical methods and skills in approximate theory,especially the profound results on width and approximate by Fourier sum.Moreover,some new ideas and skills in modern approximate theory have also played very important roles.Generally speaking,to synthesize traditional methods and modern mathematical idea and methods in functional analysis,linear algebra,probabilistic theory may provide new means to solve the problem of computational complexity of continuous problem.
其他文献
分裂可行问题(SFP)是最优化领域的重要研究课题之一,它不仅在信号处理、图像恢复上有重要的应用,而且在系统识别、经济、军事领域也有着广泛的应用.该问题自提出以来,已引起
该文来源于新华医院高干体检数据分析项目,以2000年的体检数据为研究对象,主要解决了三个问题:首先,利用因子分析及逐步Logistic回归方法对与冠心病患病关系密切的六个危险因
近几十年来,随着科学技术的发展进步,非线性问题已引起人们的广泛关注.非线性分析作为一种研究工具应运而生,并迅速成为应用数学中的重要研究方向之一.在物理,生物,动力系统等领域
带预条件的GMRES算法是用来求解大型稀疏非对称问题的一种常用方法.Gene H. Golub和Denis Vanderstraeten在文献[2]中提了一种所谓反对称分裂预条件方法,该预条件方法主要是
该篇文章在前人总结出来的投资连接型保单定价的模型上,进一步分析了当死亡效力函数μ(t)为非常数时(即死亡效力与年龄有关)几类投资连接产品的定价问题.该文一共分为三部分,
对于非线性系统,其控制问题一直是学者们研究的重点.本文利用Backstepping方法,深入的研究了两类不确定非线性系统的输出反馈扰动抑制问题.本文首先介绍了非线性系统控制问题
Boltzmann方程是气体运动论的基本方程.它不仅是研究经典气体的有力工具,而且为研究固体和等离子体中的电子输运,在核反应堆中的中子输运等提供着有效的方法.在气体运动论中,