第一章，从阿尔法零的卓越性能出发，深入解读其背后着实不易的成长历程，揭示其数学模型。第二章，从确定性和随机动态规划问题入手，介绍决策问题的数学模型。第三章，从抽象视角回顾纷繁复杂的强化学习算法，揭示值函数近似与滚动改进的重要作用。第四章，从经典的线性二次型**控制问题入手，分析从阿尔法零的成功中学到的经验。第五章，分别从鲁棒、自适应、模型预测控制等问题入手，分析值函数近似与滚动改进对算法性能的提升潜力。第六章，从离散优化的视角审视阿尔法零的成功经验。第七章，总结全书。适合作为本领域研究者作为学术专著阅读，也适合作为研究生和本科生作为参考书使用。

[美]德梅萃·P. 博塞克斯（Dimitri P. Bertseka）,美国MIT终身教授，美国国家工程院院士，清华大学复杂与网络化系统研究中心客座教授。电气工程与计算机科学领域国际知名作者，著有《非线性规划》《网络优化》《动态规划》《凸优化》《强化学习与**控制》等十几本畅销教材和专著。

Preface With four parameters I can fit an elephant, and with five I can make him wiggle his trunk4 John von Neumann The purpose of this monograph is to propose and develop a new concep- tual framework for approximate Dynamic Programming (DP) and Rein- forcement Learning (RL). This framework centers around two algorithms, which are designed largely independently of each other and operate in syn- ergy through the powerful mechanism of Newton's method. We call these the off-line training and the on-line play algorithms; the names are bor- rowed from some of the major successes of RL involving games. Primary examples are the recent (2017) AlphaZero program (which plays chess...