Fall 2002

Autonomous Agent Architecture (CMPS 522)

 

An autonomous-agent, which can be hardware, software or a hybrid of both, is capable of sensing the environment, making decisions, and performing single or a sequence of actions to the environment so as to achieve its goal(s). In this course, I will start with a discussion of different architectures proposed to build an autonomous agent and then talk about issues of knowledge representation, and machine learning. We will then discuss the issues of, and approaches for, communication and collaboration of agents in a distributed environment. 

 

The specific topics covered include

 

Multiagent Systems and Societies of Agents

Agent Communications, Agent Interaction Protocols, Societies of Agents and Negotiation                                                                              

Distributed Problem Solving and Planning           

Task Sharing, Result Sharing, Distributed Planning, Distributed Plan Representations, and Distributed Planning and Execution

                                                

Search Algorithms for Agents

Constraint Satisfaction, Path-Finding Problem, Two-Player Games

 

Distributed Rational Decision Making

Evaluation Criteria, Voting, Auctions, Bargaining, General Equilibrium Market Mechanisms, Contract Nets and Coalition Formation   

 

Learning in Multiagent Systems

A General Characterization, Learning and Activity Coordination, Reinforcement Learning, Learning about and from Other Agents, and Learning and Communication

 

Formal Methods in DAI

 

Industrial and Practical Applications of DAI

           

Distributed Models for Decision Support

 

Evaluation:

It has a midterm examination, home works and a semester project.

Prerequisite:

Previous AI knowledge or background (CMPS 420, equivalent or consent of the instructor)

Time and Place:

Meets on Tuesday and Thursday at 3:30 pm   in HLG 128

Text Books:

1.    Multiagent Systems: A modern approach to distributed artificial intelligence edited by G. Weiss (ISBN: 0262731312)

 

Reference Books

1. Artificial Intelligence by Nills J. Nilsson published by Morgan Kaufmann (chapters 2, 5, 23, 24 and 25)

2. Artificial Intelligence: A modern Approach by Stuart Russell and Peter Norvig Published by Prentice Hall (chapter 2)

3. Reading in Agents Edited by Michael N. Huhns and Munindar P. Singh, Published by Morgan Kaufmann

4. Software Agents, Edited by Jeffrey M. Bradshaw, published by The MIT Press

5. Multi Agent Systems By Jacques Ferber, Published by Addison-Wesley

6. Intelligent Software Agents by Brenner at al. Published by Springer