Introduction to Multi-Agent Systems

Author: Wardley
Publisher:
Publish Date: 2003-10-01
Features: Multi-Agent Systems (MAS) represent a new approach to conceptualizing and implementing distributed software. This book is a textbook on the important topic of Multi-Agent Systems, providing a comprehensive overview of Intelligent Agents and MAS. It introduces the concept of Agents as autonomous action-capable software systems. The book guides readers through detailed discussions on:
● Methods for building Agents
● How Agents achieve consensus
● The languages used for communication between Agents
● Interoperability and coordination in Agent systems
● Applications of Agent technology
Written for university students and computer professionals, this book is supported by online teaching resources, including a complete set of teaching slides. It features numerous informal examples for illustration and lists over 500 external references. The book systematically introduces the fundamental knowledge of Agents and MAS, summarizing research findings from the 1980s to 2001. It serves as an excellent resource for understanding Agent/MAS.
MAS belongs to the field of Distributed Artificial Intelligence within the discipline of AI. An Agent is an entity with strong expressive capabilities. For example, in describing a robot soccer match, the field, goal, kicking rules, goalkeeper, and each player can all be represented as separate Agents. If Player 1 is represented as Agent1, Agent1 should possess extensive knowledge about soccer, understand defensive players, and have abilities such as dribbling and shooting. In contrast, the Agent representing the goal is much simpler, requiring only information about its position, height, and width.
MAS is a system where multiple Agents collaborate to solve problems. Each Agent has its own goals, and the collective Agent (e.g., scoring a goal) also pursues objectives. The problem-solving process can utilize market mechanisms, game theory, and other methods to achieve cooperation.
The book is divided into 12 chapters, introducing the basic concepts of Agents and MAS, cooperative problem-solving methods, and application examples. Topics include what Agents are, their reasoning capabilities, MAS and Brooks' view of intelligence, Agent communication, cooperative problem-solving among Agents, and finally the possible-world models of Agents based on modal logic.
This book can serve as an introductory textbook for university courses on AI systems introducing Agents and MAS, as well as a reference for researchers engaged in AI research.