出版时间:2011-1 出版社:清华大学出版社 作者:博赛克斯 页数:403
Tag标签:无
前言
优化理论与应用是非常经典但依然非常活跃的研究领域,涉及几乎所有的理工和管理学科以及计量社会科学学科,是系统工程、运筹学、计量经济学等学科的理论基础。凸优化是优化理论十分重要的分支,是本书讨论的重点。凸优化是指目标函数为凸函数、约束集为凸集合的约束优化问题。凸优化具有重要的工程应用背景,求解凸优化问题的方法通常也是一般非线性规划方法的重要基础。本书是凸优化理论与方法的重要专著和教材,主要内容分为两部分:凸分析和凸问题的对偶优化理论。本书先从基本线性代数和实分析理论出发,比较详尽地讨论了凸理论和凸分析,为求解凸优化问题建立了足够的基础。本书在引入了凸优化的基本概念后,着重讨论了对偶优化理论。本书从比较独特的几何问题角度——最小共同点和最大相交点问题——引入了对偶理论框架,讨论对偶性和对偶优化解的存在性等问题。在此统一对偶理论框架下,本书讨论了多种优化问题如线性规划、凸规划、最小最大等问题的对偶性和对偶优化理论,并讨论了当目标函数非光滑时的次梯度和最优性条件。本书的重要特点是白成体系,所需要的基础知识除理工科本科线性代数和少量实分析基本概念和理论外,并不需要一般优化理论如线性规划、非线性规划等作为基础。所以本书既适用作研究生的教材,也可作为优化理论与方法研究者的参考书。本书作者德梅萃·博赛克斯教授是优化理论的国际著名学者、美国国家工程院院士,现任美国麻省理工学院电气工程与计算机科学系教授,曾在斯坦福大学工程经济系和伊利诺伊大学电气工程系任教,在优化理论、控制工程、通信工程、计算机科学等领域有丰富的科研教学经验,成果丰硕。博赛克斯教授是一位多产作者,著有14本专著和教科书。本书是作者在优化理论与方法的系列专著和教科书中的一本,自成体系又相互对应。
内容概要
本书作者德梅萃,博赛克斯教授是优化理论的国际著名学者、美国国家工程院院士,现任美国麻省理工学院电气工程与计算机科学系教授,曾在斯坦福大学工程经济系和伊利诺伊大学电气工程系任教,在优化理论、控制工程、通信工程、计算机科学等领域有丰富的科研教学经验,成果丰硕。博赛克斯教授是一位多产作者,著有14本专著和教科书。本书是作者在优化理论与方法的系列专著和教科书中的一本,自成体系又相互对应。主要内容分为两部分:凸分析和凸问题的对偶优化理论。
作者简介
作者:(美国)博赛克斯(Dimitri P.Bertsekas)
书籍目录
1. basic concepts of convex analysis 1.1. convex sets and functions 1.1.1. convex functions 1.1.2. closedness and semicontinuity 1.1.3. operations with convex functions 1.1.4. characterizations of differentiable convex functions 1.2. convex and afiine hulls 1.3. relative interior and closure 1.3.1. calculus of relative interiors and closures 1.3.2. continuity of convex functions 1.3.3. closures of functions 1.4. recession cones 1.4.1. directions of recession of a convex function 1.4.2. nonemptiness of intersections of closed sets 1.4.3. closedness under linear transformations 1.5. hyperplanes 1.5.1. hyperplane separation 1.5.2. proper hyperplane separation 1.5.3. nonvertical hyperplane separation 1.6. conjugate functions 1.7. summary 2. basic concepts of polyhedral convexity 2.1. extreme points 2.2. polar cones 2.3. polyhedral sets and functions 2.3.1. polyhedral cones and farkas' lemma 2.3.2. structure of polyhedral sets 2.3.3. polyhedral functions 2.4. polyhedral aspects of optimization 3. basic concepts of convex optimization 3.1. constrained optimization 3.2. existence of optimal solutions 3.3. partial minimization of convex functions 3.4. saddle point and minimax theory 4. geometric duality framework 4.1. min common/max crossing duality 4.2. some special cases 4.2.1. connection to conjugate convex functions 4.2.2. general optimization duality 4.2.3. optimization with inequality constraints 4.2.4. augmented lagrangian duality 4.2.5. minimax problems 4.3. strong duality theorem 4.4. existence of dual optimal solutions 4.5. duality and polyhedral convexity 4.6. summary 5. duality and optimization 5.1. nonlinear farkas' lemma 5.2. linear programming duality 5.3. convex programming duality 5.3.1. strong duality theorem inequality constraints 5.3.2. optimality conditions 5.3.3. partially polyhedral constraints 5.3.4. duality and existence of optimal primal solutions 5.3.5. fenchel duality 5.3.6. conic duality 5.4. subgradients and optimality conditions 5.4.1. subgradients of conjugate functions 5.4.2. subdifferential calculus 5.4.3. optimality conditions 5.4.4. directional derivatives 5.5. minimax theory 5.5.1. minimax duality theorems 5.5.2. saddle point theorems 5.6. theorems of the alternative 5.7. nonconvex problems 5.7.1. duality gap in separable problems 5.7.2. duality gap in minimax problems appendix a: mathematical background notes and sources supplementary chapter 6 on convex optimization algorithm
章节摘录
插图:Convex sets and functions are very useful in optimization models, and havea rich structure that is convenient for analysis and algorithms. Much of thisstructure can be traced to a few fundamental properties. For example, eachclosed convex set can be described in terms of the hyperplanes that supportthe set, each point on the boundary of a convex set can be approachedthrough the relative interior of the set, and each halfline belonging to aclosed convex set still belongs to the set when translated to start at anypoint in the set.Yet, despite their favorable structure, convex sets and their analysisare not free of anomalies and exceptional behavior, which cause seriousdifficulties in theory and applications. For example, contrary to affineand compact sets, some basic operations such as linear transformation andvector sum may not preserve the closedness of closed convex sets. This inturn complicates the treatment of some fundamental optimization issues,including the existence of optimal solutions and duality.For this reason, it is important to be rigorous in the development ofconvexity theory and its applications. Our aim in this first chapter is toestablish the foundations for this development, with a special emphasis onissues that are relevant to optimization.
编辑推荐
《凸优化理论(影印版)》:清华版双语教学用书
图书封面
图书标签Tags
无
评论、评分、阅读与下载