Universidade Nova de Lisboa

Universidade Nova de Lisboa

Faculdade de Ciências e Tecnologia

Multi-Agent Systems

Code

8292

Academic unit

Faculdade de Ciências e Tecnologia

Department

Departamento de Informática

Credits

6.0

Teacher in charge

João Alexandre Carvalho Pinheiro Leite, José Júlio Alves Alferes

Weekly hours

4

Total hours

56

Teaching language

Português

Objectives

Knowledge:

  • Agent Architectures
  • The Agent Oriented Programming Lnaguage Jason
  • Mechanisms and Strategies for individual decision making in Multi-Agent Systems
  • Mechanisms and Strategies for negotiation, distributed decision making and colaboration in Multi-Agent Systems.

Know-how:

  • Analyse and model a problema using a multi-agent system
  • Implement a multi-agent system in Jason

Subject matter

  1. Introduction to Multi-Agent Systems
  2. Agent Architectures
    1. Intelligent Agents
    2. Deductive Reasoning Agents
    3. Practical Reasoning
    4. Reactive and Hybrid Architectures
  3. Agent-oriented Programming Languages
  4. Non-cooperative game theory
  5. Social Choice Theory
  6. Bayesian Games
  7. Mechanism Design

Bibliography

M. Wooldridge. An Introduction to Multiagent Systems. 2nd Edition, John Wiley, 2009.

Y. Shoham and K. Leyton-Brown, Multiagent Systems: Algorithmic, Game-theoretic, and Logical Foundations. Cambridge University Press, 2009.

Teaching method

The syllabus is taught in theoretical and laboratory classes. In the former, the main concepts and techniques are addressed.

The laboratory classes are dedicated to solving problems, and experimenting the various concepts of the syllabus using the Jason, and (partially) developing the project.

Evaluation method

The evaluation includes:

  • theoretical componente: 2 midterm individual tests (or an exam) where the knowledge acquired by the students on the concepts and techniques is assessed.
  • practical component: 1 practical project, developed in groups of 2 students, consisting of the design, analysis and implementation of a team of agents to participate in the agent programming competition (http://multiagentcontest.org).

The project is developed/evaluated in two steps: first the design of the multi-agent system; second, the complete project. The evaluation of the final project includes a competition between all submitted systems.

For passing in the course a student needs a mark in the theoretical part equal or above 10. The final grade is the average of both components.

Courses