Preamble

In this course on Multi-Agent Programming, the focus is on the practical programming of Multi-Agent Systems. It address one of the main lines of use of the multi-agent paradigm for developing complex systems: Multi-Agent Oriented Programming (MAOP) used for the integration of intelligent systems and development of socio-technical systems.

Besides learning foundations and state of the art works done on this topic in the Multi-Agent domain, you will learn how to take a MAOP approach for developing Decentralized and Open AI systems.

The Multi-Agent Oriented Programming approach consists in a composition of Agent models, Environment models and Organisation models, interacting all together. Each of these models are built by picking concepts in, respectively, the Agent, Environment or Organisation dimension. Using MAOP by privileging each of these dimensions may respectively lead to Agent Oriented Programming (AOP), Environment Oriented Programming (EOP) and Organization Oriented Programming (OOP).

We use the JaCaMo platform for the practical exercises that are presented in this course. This programming platform results from the integration of the Jason Agent Programming Language, CArTaGo Environment Programming platform and MOISE Organization programming platform.

Course Organisation

Multi-Agent Oriented Programming

Introduction to Multi-Agent Systems and to Multi-Agent Oriented Programming

Programming Agents

Programming Environment

Programming Organization and norms

MAOP approach on coordination

Working Environment Configuration (For JaCaMo users)

  • get a release of the JaCaMo platform here

  • install the eclipse plugin as described here or configure your shell command as described here

  • more documentation:

Multi-Agent based simulation

  • Introduction to Multi-Agent Based Simulation

  • Simulation coupling

  • REPAST

Project

  • The project is described here with the agenda and deliverables to produce