CENG 580 - Multiagent Systems
Instructor: Faruk Polat
Catalog Data: Multigent Systems (3-0) 3
Concurrency and distribution in AI. Agents:
micro and macro views. Agent communication languages.
Rational agency: economic/game theoretic, logical. BDI
architecture. Multi-agent real-time
search. Reinforcement learning. Multi-agent learning.
Opponent modeling.
Prerequisites: Consent of the instructor.
- Textbook(s): None
- Reference(s):
- G.Weiss, Multi-Agent Systems: A Modern Approach to Distributed Artificial Intelligence, MIT Press, 2000.
- Y.Shoham and K. Leyton-Brown, Multiagent Systems: Algorithmic, Game-Theoretic and Logical Foundations, Cambridge University Press, 2009.
- M.Wooldridge, An Int. to MultiAgent Systems, John Wiley & Sons, 2002.
- Selected classics in Autonomous Agents and Multiagents Systems Journal, Artificial Intelligence Journal, Journal of Artificial Intelligence Research and several conferences including AAMAS, IJCAI, AAAI..
- Goals: To develop the theory and practice of multi-agent systems.
- Course Outline:
Topics
- Introduction to autonomous agents and multi-agent systems
- Purely reactive agents
- Perception
- Agents with state
- Concrete architectures for intelligent agents
- Logic based architectures
- Reactive architectures
- Belief, desire, intention architectures
- Layered architectures
- Agent Communication
- Speech Act Theory
- Agent Communication Languages: KQML and ACL
- Real-Time Search
- Real-time A* (RTA*),
- LRTA (Learning Real-Time A*),
- Moving Target Search (MTS)
- Multiagent Search
- Multi-agent Real-Time A* (MARTA*)
- Organizational strategies
- Rational Agents
- Agents as rational decision makers
- Observable worlds and Markov Property
- Stochastic transitions and utilities
- Game theory
- Agent Interaction
- Strategic games
- Iterated elimination of strictly dominated actions
- Nash equilibrium
- Mechanism design, negotiation
- Project Progress Presentation
- Logic-based agency
- Modal logic, dynamic logic, temporal logic
- BDI architectures
- Reinforcement Learning
- TD(λ), Q-learning
- Learning in Markov Games
- NASH Q-Learning, Mini-Max Q-Learning
- Action Estimation Algorithm (ACE) and Action Group Estimation Algorithm (AGE)
- Research Paper Presentations
Important Dates
- Project proposal report.
- Project progress presentations
- Project progress report.
- Project Report and Demo :
Evaluation:
- Course Project: (progress report 15%, final report+demo 35%)
- Homeworks: Programming Assignments. 25%
- Research Paper Presentations (15%): technical presentation.
- Active participation in the presentations (10%):
Homeworks: Individual programming tasks
Grades: