The course Multi-agent systems deals with development of systems with a specific kind of modularity based on agent metaphor.

MAS 2010

1. Multiagent system example. Agent metaphor. Complex system. Robustness. Decentralisation. Artificial Intelligence. Weak agents and emergence. Strong agents and AI employment. Multi-agent system. Agents classification. Communication among agents. lecture 1

2. Representation and communication language. lecture 2 (play with XSLT)

3. Representation and Commnunication Language. Implementation of MAS within network programming, OOP and IPC. Prerequisity: Transport layer of ISO OSI. Middleware. lecture 3

4. MAS as a kind of middleware. Programming on middle tier: RMI, RMI-IIOP, CORBA, WebServices(Axis). lecture 4

5. MAS as a kind of middleware. Direct communication. Programming with JADE. FIPA, ACL. AMS, DF. lecture 5

6. MAS as a kind of middleware. Indirect communication. Client-Server. Agent-Space. lecture 6

7. MAS over OO virtual machine. Multithreading. Synchronization. lecture 7

8. MAS over IPC. SRR model. Master-Slave. Client-Server. Agent-Space. lecture 8

9. MAS based control. Subsumption. lecture 9

10. Behavioral architectures. MAS-based modelling of biological systems. Failures and delusions supporting evidence of MAS organisation of living systems. lecture 10

11. Agent-Space architecture. Modelling in real-time. lecture 11

12. Incorporation of traditional AI approach to MAS. lecture 12

Training:
training 1. Prerequsities: Java
training 2. Prerequsities: VRML: statics and dynamics (Play with VRML)
training 3. Prerequsities: example of VRML scene controlled via TCP/IP: bullet collision (programming on communication tier)
training 4. Simple space implemented via RMI or CORBA (programming on middle tier)
training 5. Direct communication with JADE
training 6. Indirect communication with Agent-Space: Remote control of bullet in VRML scene via simple MAS implemented over RMI
training 7. Proprietary multi-agent system over multithreading of JVM
training 8. Programming over SRR under JavaShell
training 9. Multi-agent Control. Reimplementation of mobile robot ALLEN
training 10. Behavioral architecture for LEGO robot
training 11. Robot following ball controlled by set of heterogeneous agents
training 12. Inteligent agent controlled by reinforcement learning

Students ranking
1.Michal Vince18
2.Michal Stríženec17
3.Miroslav Medveď15
4.Ondrej Mikuláš14.5
Michal Gašpierik14.5
6.Jaroslav Blanář13.5
7.Petra Horňáková13
8.Lukáš Bača 12

Skúška bude 26.5. 9:00 v niektorom akváriu

Photos & videos: Students 2006
A video from summer term of 2006/2007 (submitted by Juraj Bednar)
Students 2010

Requirement:
almost 100% presence at tranings or project
ability to solve all problems demonstrated on training

Literature:
mainly lectures and my publications
Kelemen J.: Strojovia a agenty
Kubik. A.: Agenty a multiagentove systemy

This course is part of educational program at http://www.ii.fmph.uniba.sk/

 

Others:

Machine vision for young programmers: pdf ppt Skflag.gif (756 bytes)
Evolution of programming structures pdf ppt Skflag.gif (756 bytes)
Embodiment from Engineer's Point of View pdf ppt Ukflag.gif (756 bytes)
Hidden Markov Models pdf ppt Skflag.gif (756 bytes)
Principal component analysis pdf ppt Skflag.gif (756 bytes)
Subsumption architecture pdf ppt video Ukflag.gif (756 bytes) pdf ppt video Skflag.gif (756 bytes)
How can computer recognize orange color? pdf ppt Skflag.gif (756 bytes)
Color vision pdf Skflag.gif (756 bytes)
Introduction to robot vision pdf ppt training Ukflag.gif (756 bytes)
Object mapping (lecture for course "Tvorba informačných systémov") pdf ppt Ukflag.gif (756 bytes)
Object Recognition: Histogram of oriented gradients (HOG) & Dominant orientation templates (DOT) (by Michal Vician) pdf Skflag.gif (756 bytes)
LEGO robots
Star from computer tape (KSP, Demacek)

Diploma thesis:

2007/08 Robosoccer by Milan Leitman video