Artificial Intelligence 2

The course is a substitute for:

Artificial Intelligence 2 (33UI2)

Lecturer:

Tutor:

Supervisor:

Department of Cybernetics

Synopsis:

The core topic of this course is building distributed systems as communities of autonomous intelligent agents. Methods and techniques for communication, coordination and cooperation among agents will be presented. The methods of team and coalition formation and auction mechanisms for distributed decision making and planning in environments with different complexity will be studied. Lecturer will explain formal models of shared mental states (BDI) and present standard referential architectures (FIPA) of implemented multi-agent systems. Algorithmic complexity of the agents' decision making will be analyzed in the end of the course.

Requirements:

Attending the labs, completing the Task 1 and the Task 2, demonstrating and presenting the results

Syllabus of lectures:

1. Concept of rationality in artificial intelligence, bounded rationality,

calculative rationality

2. Planning in artificial intelligence, planning problem representation, linear and non-linear planning

3. Planning methods: TOPLAN, POPLAN, SATPLAN, GRAPHPLAN

4. Introduction to multi-agent systems and distributed artificial intelligence, basic agent properties

5. Industrial applications of multi-agent systems

6. Reactive agents, collective reactive intelligence

7. Deliberative agents, modeling agents mental attitudes (BDI architecture) 8. Formal methods for reasoning about rational agents

9. Distributed decision making, negotiation techniques

10.Contract-net-protocol, auction techniques, voting in computational systems

11.Foundations of the game theory, prisoners dilemma

12.Coalition formation and teamwork, coordination among the agents

13.Socially oriented reasoning, social intelligence, meta-reasoning and meta-agents

14.Philosophy of artificial intelligence, conclusive remarks

Syllabus of tutorials:

1. Introduction to Lisp, CLOS (Common Lisp Object System)

2. State space search in Lisp, Task 1 - multi-agent system model

3. Task 1 - multi-agent system model

4. Task 1 - multi-agent system model

5. Demonstration of multi-agent systems ProPlanT, CPlanT and ExPlanTech

6. Open Source implementation of FIPA architectures - FIPA-OS, JADE, ZEUS

7. JADE (Java Agent DEvelopment Framework) environment tutorial, building an agent

8. Inter-agent communication in JADE

9. Design of coordination protocols in JESS (Java Expert System Shell)

10. Task 2 - implementing distributed decision making algorithms in JADE

11. Task 2 - implementing distributed decision making algorithms in JADE

12. Task 2 - implementing distributed decision making algorithms in JADE

13. Task 2 - implementing distributed decision making algorithms in JADE

14. Demonstration and results presentation

Study Objective:

Study materials:

[1] FIPA - www.fipa.org, www.emorphia.com

[2] JADE - sharon.cselt.it/projects/jade/

[3] Wooldridge M., Jennings N.: Intelligent Agents: Theory and Practice. The Knowledge Engineering Review, 10 (1995), No.2, pp. 115-1526.

Note:

Further information:

No time-table has been prepared for this course

The course is a part of the following study plans:

• Cybernetics and Measurements - Artificial Intelligence- structured studies (compulsory course)