Optimization: Principles and Algorithms
Optimization: Principles and Algorithms
复制标题
优化：原理和算法
DOI：
--
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
Bierlaire Michel
中科院分区：
文献类型：
--
作者：
Bierlaire Michel
每一位工程师和决策科学家都必须精通优化，这是他们工具包中的一个基本要素。因此，这本清晰的介绍性教科书肯定会受到学生和实践专业人士的欢迎。从他丰富的教学经验，作者巧妙地引导读者通过丰富的主题选择在一个连贯的，流体和品味的混合模型和方法锚定的基本数学概念（只有先决条件：第一年微积分和线性代数）。主题范围从经典到光滑无约束和约束优化的一些最新发展，如下降法，共轭梯度，牛顿和拟牛顿方法，线性规划和单纯形法，信赖域和内点方法。此外，离散和组合优化的元素，如网络优化，整数规划和启发式局部搜索方法也被提出。这本书介绍了优化作为一种建模工具，除了支持问题的制定加上设计和实现有效的算法，也是一种适合跨学科的人类互动的语言。读者进一步意识到，虽然数学优化的根源可以追溯到像牛顿，拉格朗日，柯西，欧拉或高斯这样的巨人的工作，但它直到第二次世界大战才成为一门学科。而且，它目前的势头实际上是由于它与现代计算机的共生关系，这使得它有可能经常解决数百万变量和约束的问题。米歇尔·比莱尔以其诙谐、有趣而又精确的风格吸引了他的读者，并唤醒了他们以创造性的方式尝试所呈现材料的欲望。本书的一个突出优点是对各种算法的统一、清晰和简洁的呈现，这使得它们易于阅读并可翻译成任何高级编程语言。“这是一本令人上瘾的书，我很高兴推荐。托马斯教授M.利布林
Every engineer and decision scientist must have a good mastery of optimization, an essential element in their toolkit. Thus, this articulate introductory textbook will certainly be welcomed by students and practicing professionals alike. Drawing from his vast teaching experience, the author skillfully leads the reader through a rich choice of topics in a coherent, fluid and tasteful blend of models and methods anchored on the underlying mathematical notions (only prerequisites: first year calculus and linear algebra). Topics range from the classics to some of the most recent developments in smooth unconstrained and constrained optimization, like descent methods, conjugate gradients, Newton and quasi-Newton methods, linear programming and the simplex method, trust region and interior point methods. Furthermore elements of discrete and combinatorial optimization like network optimization, integer programming and heuristic local search methods are also presented. This book presents optimization as a modeling tool that beyond supporting problem formulation plus design and implementation of efficient algorithms, also is a language suited for interdisciplinary human interaction. Readers further become aware that while the roots of mathematical optimization go back to the work of giants like Newton, Lagrange, Cauchy, Euler or Gauss, it did not become a discipline on its own until World War Two. Also that its present momentum really resulted from its symbiosis with modern computers, which made it possible to routinely solve problems with millions of variables and constraints. With his witty, entertaining, yet precise style, Michel Bierlaire captivates his readers and awakens their desire to try out the presented material in a creative mode. One of the outstanding assets of this book is the unified, clear and concise rendering of the various algorithms, which makes them easily readable and translatable into any high level programming language. 'This is an addictive book that I am very pleased to recommend.' Prof. Thomas M. Liebling