All programmes > Multi-Agent Systems

Multi-Agent Systems (B-KUL-H02H4A)

4 ECTSEnglish32 By way of exception, this course will not be organised this academic yearNot organisedCannot be taken as part of an examination contract
N.
POC Artificial Intelligence

Aims

This course studies the research area of multi-agent systems.  Multi-agent systems are particularly interesting for modeling and developing a wide range of distributed applications, including internet applications, distributed control systems, robotics, and self-managing systems.

In particular, the course aims to:
 
- provide an introduction and overview of software models and techniques for multi-agent systems – for behavior decision making, planning and coordination;
- provide a general perspective on the domain of collective and cooperative behaviour;
- provide a conceptual framework for distributed problem solving, based on recent research in this area.

 

Previous knowledge

  • Basic concepts and techniques on AI
  • Object-oriented programming (Java, C++, …) – for programming exercises and task.

Is included in these courses of study

Activities

3 ects. Multi-Agent Systems: Lecture (B-KUL-H02H4a)

3 ECTSEnglishFormat: Lecture24 By way of exception, this course will not be organised this academic yearNot organised
N.
POC Artificial Intelligence

Content

Introduction

  • Agents and autonomy, agents vs. objects, agents vs. expert systems
  • Cooperative vs competitive multi-agent systems
  • Distributed vs centralised vs decentralised
  • Task- vs state- vs worth-oriented domains
  • Application domains
    • cyberphysical systems (CPS) - robotics, logistics, automated driving
    • agent-based modelling
  • Agent, multi-agent systems and software engineering

 

Autonomous agents & agent architectures

  • deliberative / theoretical reasoning agents
  • reactive and behaviour-based agents, Brooks’ subsumption architecture, Agent network architecture
  • practical reasoning agents - belief-desire-intention (BDI)
  • horizontally and vertically layered architectures

 

Automated planning & acting

  • Planning vs acting
    • descriptive vs operational models
  • Planning: models, properties & algorithms
    • Deterministic / classical planning
    • Temporal planning
    • Non-deterministic planning & discretization
    • Probabilistic planning
    • Hierarchical task networks
  • PDDL

 

Multi-agent planning

  • POCL planning
  • Parallel POCL planning
  • Multi-agent POCL planning
  • MA-PDDL

 

Multi-agent task allocation

  • contract net + variants (incl dynamic, with confirmation)
  • task trees and subtree bidding
  • gradient fields
  • agent negotiations for task re-distribution
    • deals, conflict deal, Zeuthen strategy

 

Swarm intelligence

  • emergent behaviour, self-organization
  • stigmergy
  • ACO (ant colony optimisation) - TSP, AntNet, AntSystem

 

Delegate MAS for large-scale dynamic coordination and control applications

  • task & resource agents
  • exploration, intention and feasibility
  • case study - manufacturing control & logistics

 

Competitive multi-agent systems & game theory

  • non-cooperative game theory
  • coalitional game theory

Course material

  • An Introduction to MultiAgent Systems - Second Edition, M. Wooldridge, 2009.
  • M. Ghallab, D. Nau, P. Traverso, “Automated planning and acting” - http://projects.laas.fr/planning/, 2016
  • Various papers and book chapters from literature.

1 ects. Multi-Agent Systems: Project (B-KUL-H08M2a)

1 ECTSEnglishFormat: Assignment8 By way of exception, this course will not be organised this academic yearNot organised
N.
POC Artificial Intelligence

Content

A scientific project aims at experiencing the challenges as well as the opportunities that multi-agent systems entail in distributed problem solving. The project includes a limited literature study, practical development, evaluation, reporting.

Evaluation

Evaluation: Multi-Agent Systems (B-KUL-H22H4a)

Type : Partial or continuous assessment with (final) exam during the examination period
Description of evaluation : Written, Paper/Project
Type of questions : Closed questions, Open questions

Explanation


  • Questions & exercises on the covered concepts and techniques for MAS-based modeling and problem solving
  • Discussion of project work