多智能体系统可以看作由多个具有自主决策能力的软件智能体组成，各智能体之间会直接或间接地相互作用和影响。通常可以把多智能体系统分为两大类：合作式多智能体系统和非合作式多智能体系统，前者研究的核心问题是各智能体如何利用有限的局部信息，通过自主学习有效协作达到优化的共同目标；而后者一个重要问题是如何采用有效激励机制，促使各智能体主动协调合作，从而提升系统的整体性能。郝建业、梁浩锋著的本书将涵盖以公平性等为目标的多智能体协调合作理论与技术，结合不同优化目标，介绍新的多智能体学习算法和激励机制研究进展。本书适用于对多智能体系统设计理论感兴趣的读者，也可作为从事多智能体系统及博弈论理论研究的研究生或科研人员的参考书籍，对从事多智能体系统软件开发人员也具有一定的参考价值。

1  Introduction
1.1  Overview of the Chapters
1.2  Guide to the Book
References
2  Background and Previous Work
2.1  Background
2.1.1  Single-Shot Normal-Form Game
2.1.2  Repeated Games
2.2  Cooperative Multiagent Systems
2.2.1  Achieving Nash Equilibrium
2.2.2  Achieving Fairness
2.2.3  Achieving Social Optimality
2.3  Competitive Multiagent Systems
2.3.1  Achieving Nash Equilibrium
2.3.2  Maximizing Individual Benefits
2.3.3  Achieving Pareto-Optimality
References
3  Fairness in Cooperative Multiagent Systems
3.1  An Adaptive Periodic Strategy for Achieving Fairness
3.1.1  Motivation
3.1.2  Problem Specification
3.1.3  An Adaptive Periodic Strategy
3.1.4  Properties of the Adaptive Strategy
3.1.5  Experimental Evaluations
3.2  Game-Theoretic Fairness Models
3.2.1  Incorporating Fairness into Agent Interactions
Modeled as Two-Player Normal-Form Games
3.2.2  Incorporating Fairness into Infinitely Repeated
Games with Conflicting Interests for Conflict Elimination
References
4  Social Optimality in Cooperative Multiagent Systems
4.1  Reinforcement Social Learning of Coordination
in Cooperative Games
4.1.1  Social Learning Framework
4.1.2  Experimental Evaluations
4.2  Reinforcement Social Learning of Coordination
in General-Sum Games
4.2.1  Social Learning Framework
4.2.2  Analysis of the Learning Performance Under
the Social Learning Framework
4.2.3  Experimental Evaluations
4.3  Achieving Socially Optimal Allocations Through Negotiation
4.3.1  Multiagent Resource Allocation Problem
Through Negotiation
4.3.2  The APSOPA Protocol to Reach Socially Optimal
Allocation
4.3.3  Convergence of APSOPA to Socially Optimal Allocation..
4.3.4  Experimental Evaluation
References
5  Individual Rationality in Competitive Multiagent Systems
5.1  Introduction
5.2  Negotiation Model
5.3  ABiNeS: An Adaptive Bilateral Negotiating Strategy
5.3.1  Acceptance-Threshold (AT) Component
5.3.2  Next-Bid (NB) Component
5.3.3  Acceptance-Condition (AC) Component
5.3.4  Termination-Condition (TC) Component
5.4  Experimental Simulations and Evaluations
5.4.1  Experimental Settings
5.4.2  Experimental Results and Analysis: Efficiency
5.4.3  Detailed Analysis of ABiNeS Strategy
5.4.4  The Empirical Game-Theoretic Analysis: Robustness
5.5  Conclusion
References
6  Social Optimality in Competitive Multiagent Systems
6.1  Achieving Socially Optimal Solutions in the Context
of Infinitely Repeated Games
6.1.1  Learning Environment and Goal
6.1.2  TaFSO: A Learning Approach Toward SOSNE Outcomes:
6.1.3  Experimental Simulations
6.2  Achieving Socially Optimal Solutions in the Social
Learning Framework
6.2.1  Social Learning Environment and Goal
6.2.2  Learning Framework
6.2.3  Experimental Simulations
References
7  Conclusion
Reference
A The 57 Structurally Distinct Games