Quanyan Zhu

A graduate textbook on agents that plan, call tools, and act with delegated authority, built on decision theory, games, and learning. Three parts cover agents as decision-makers, building agentic systems, and security, resilience, and governance.

Sample: preface, full contents, and Part I (Chapters 1–3), 126 pages.

How agents learn in strategic environments: behavioral, cognitive, and epistemic learning, from equilibrium and non-equilibrium dynamics to world models and conjectural learning. Fifteen chapters in five parts, about 570 pages.

Sample: contents, notation, and Chapters 1–3, 121 pages.

The economics of the token as the unit through which machine intelligence is produced, metered, priced, and governed. Twenty-one chapters and four appendices, about 710 pages.

Sample: preface, full contents, and Part I (Chapters 1–3), 103 pages.

A quantitative treatment of space systems as contested critical infrastructure, spanning signal, link, and network security, physical threats and orbital games, cyber-physical security, and resilient architecture and governance. Twenty chapters in six parts and six appendices, about 1,080 pages.

Sample: front matter, full contents, and Part I (Chapters 1–4), 192 pages.

A unified course from the mathematical foundations of optimization through convex optimization and algorithms to deterministic and stochastic optimal control, including the calculus of variations, estimation theory, and Kalman filtering. Fifteen chapters in five parts and two appendices.

Full manuscript, 789 pages.

Human and machine cognition as a security frontier, moving past “the human is the weakest link” to vulnerabilities that can be modeled and engineered against: cognitive attack chains, defenses for individuals, organizations, and societies, deception turned to defensive advantage, and humans and AI in the control loop. Twenty-two chapters in seven parts with two appendices, about 810 pages.

Sample: preface, notation, full contents, and Part I (Chapters 1–4), 156 pages.

Systems engineering rewritten for components that behave stochastically: the five classical assumptions AI violates, architecture of agents and harnesses, an adapted lifecycle from concept through verification to sustainment, dependability and security, and the economics of running such systems. Sixteen chapters in six parts, about 400 pages.

Sample: contents, Part I (Chapters 1–3), and Chapter 4, 119 pages.

A graduate course from matrix games and Nash equilibrium through continuous-kernel and Stackelberg games, incomplete information, auctions and mechanism design, dynamic, repeated, Markov, differential, stochastic and mean-field games, to cooperative, matching, routing, evolutionary and learning games. Thirty chapters in seven parts with two appendices, about 950 pages.

Sample: Part I, Fundamentals of Game Theory (Chapters 1–8), 262 pages.

A primer on market microstructure and optimal trading for readers who know stochastic control, convex analysis, and mathematical programming but not finance. A market is treated as a controlled dynamical system whose noise is endogenous, whose plant reacts to your control, and whose objective is a risk-adjusted terminal cash position: the limit order book, optimal execution, market making, hedging, portfolio selection, risk measures, backtesting, high-frequency and agentic trading. Sixteen chapters in three parts with appendices.

Full notes, 208 pages.