« SMA/old » : différence entre les versions

De Parcours SIIA
Aller à la navigation Aller à la recherche
Aucun résumé des modifications
Aucun résumé des modifications
 
Ligne 4 : Ligne 4 :
By Pierre Chevaillier, Professor at ENIB.
By Pierre Chevaillier, Professor at ENIB.


=== Part One: Basic concepts, agent architectures and technologies ===
=== Annales d'examen ===
*[[Media:Examen_2017-2018.pdf| Examen de 2017-2018]] : attention, celui-ci dure une heure, contrairement à celui de cette année (2h), qui comportera une partie portant sur le cours de de Pierre Chevaillier.
 
=== Projets 2020 - 20 ===
 
* [[Media:Projets_SMA.pdf|Présentation des projets]]
* [[Media:Projet_JR1.zip|Supports du projet JR1]]
* [[Media:Projet_JR2.zip|Supports du projet JR2]]
 
Projets proposés par Pierre Chevaillier
 
Simulation de déplacements collectifs de piétions en utilisant
* PC1 : [[Media:Siia-sma_sujet-projet_modeles_pietons_reynolds_2019.pdf|le modèle des boids de Reynolds]]
* PC2 : [[Media:Siia-sma_sujet-projet_modeles_pietons_helbing_2019.pdf|le modèle des social forces de Helbing et al.]]
* PC3 : [[Media:Siia-sma_sujet-projet_modeles_pietons_fajen-warren_2019.pdf|le modèle de la dynamique comportementale de Fajen & Warren]]
 
Les 3 modèles sont décrits dans le document [[Media:Siia-sma_steeringModels_2019.pdf|Steering models]].
 
Les scénarios de test sont ceux proposés  [[Media:Singh_etal_cavw09.pdf|l'article de Singh et al (2009)]].
 
*  PC4 : Résolution itérative de contraintes spatiales par un système multi-agents en s'inspirant de la proposition de [[Media:Agents_stresses_jfsma03.pdf|l'article de De Loor et al (2003)]].
 
=== Introduction to MultiAgent Systems ===
 
Download the [https://www.enib.fr/~chevaill/documents/master/siia-mas_multiagent-systems.pdf Course material].
 
* Part One: Basic concepts, agent architectures and technologies


This first lesson presents the basic principle related to the concepts of agent and multiagent systems. Next it gives a broad overview of the different conceptual solutions and types of technology used to designing agent-oriented software.
This first lesson presents the basic principle related to the concepts of agent and multiagent systems. Next it gives a broad overview of the different conceptual solutions and types of technology used to designing agent-oriented software.


=== Part Two: collective behavior in multi agent systems ===
* Part Two: collective behavior in multi agent systems


=== Part Three: Approches in the modeling of multiagent systems ===
* Part Three: Approches in the modeling of multiagent systems


=== IT Laboratory ===
=== IT Laboratory ===
Ligne 22 : Ligne 48 :
==== Agents using the protocol ====
==== Agents using the protocol ====


At this point, the agent that initiates the protocole (the initiator) does it only once. We want now to make the agents able to use the protocole on purpose, namely when they have something to do and cannot perform it by their own.
At this point, the agent that initiates the protocol (the initiator) does it only once. We want now to make the agents able to use the protocol on purpose, namely when they have something to do and cannot perform it by their own.


==== Many agents ====
==== Many agents ====


Implement a solution where many agents can run the protocole concurrently.
Implement a solution where many agents can run the protocol concurrently.


== Complex systems & Agent-Based Model And Simulation ==
== Complex systems & Agent-Based Model And Simulation ==
Jeremy Riviere, Associate Professor at UBO.
Jeremy Riviere, Associate Professor at UBO.


[[Media:SMA_Riviere.pdf|Complex systems, Agent-Based Models and Simulations]]
The course material is [[Media:Multi-agents_systems_and_interactive_simulation.pdf|here]]
 
=== Part One: Complex systems, Models and Simulations ===
[[Media:Systèmes_Complexes1.pdf|Introducing]] the concept of complex systems, and answering the questions "why" and "how" do we need to model and simulate them.
 
=== Part Two: MAS for modelling and simulating complex systems ===
* [[Media:SMA_pour_la_modélisation_et_la_simulation_de_systèmes_complexes,_implémentation_en_Java.pdf|Model engineering]] and implementing MAS simulation with Java.
* [[Media:TPs_et_tutoriel_Repast.zip|Practical works and tutorial]]
* [[Media:La_plateforme_JADE_Bibliographie_et_références_du_cours.pdf|JADE platform and bibliography]]

Dernière version du 17 juillet 2020 à 08:35

Multi-Agent Systems

By Pierre Chevaillier, Professor at ENIB.

Annales d'examen

  • Examen de 2017-2018 : attention, celui-ci dure une heure, contrairement à celui de cette année (2h), qui comportera une partie portant sur le cours de de Pierre Chevaillier.

Projets 2020 - 20

Projets proposés par Pierre Chevaillier

Simulation de déplacements collectifs de piétions en utilisant

Les 3 modèles sont décrits dans le document Steering models.

Les scénarios de test sont ceux proposés l'article de Singh et al (2009).

Introduction to MultiAgent Systems

Download the Course material.

  • Part One: Basic concepts, agent architectures and technologies

This first lesson presents the basic principle related to the concepts of agent and multiagent systems. Next it gives a broad overview of the different conceptual solutions and types of technology used to designing agent-oriented software.

  • Part Two: collective behavior in multi agent systems
  • Part Three: Approches in the modeling of multiagent systems

IT Laboratory

The objective of this Lab work is to implement a Contract Net Protocol (CNP). It will be implemented on the Gama platform. Here is a basic example that illustrates the facilities the Gama platform provides to implement FIPA compliant communication protocols.

The protocole

Step by step implementation of the CNP. The objective is to implement the way agents produce and process the different types of communicative acts that support the protocol.

Agents using the protocol

At this point, the agent that initiates the protocol (the initiator) does it only once. We want now to make the agents able to use the protocol on purpose, namely when they have something to do and cannot perform it by their own.

Many agents

Implement a solution where many agents can run the protocol concurrently.

Complex systems & Agent-Based Model And Simulation

Jeremy Riviere, Associate Professor at UBO.

The course material is here