Posts tagged "simulation"

Model Self-Calibration using Self-Adaptive Multi-Agent System

Context Presentation

The purpose of this project is to propose a cooperative agent model, based on the self-adaptive multi-agent system theory (AMAS), allowing an efficient and fast exploration of the parameter space, autonomously and automatically. This exploration should allow a continuous readjustment of the simulation until convergence, improving the control of the macro-level over the micro-level.

On an application standpoint, the purpose of this project is to produce a realistic traffic that satisfies the best a set of objectives and constraints at both micro and macro levels. This traffic should also allow interaction with humans and adapt to events that could occur in the virtual environment. 

CALICOBA_simple - Darmo


Self-adaptive Multi-agent Systems, Self-Calibration, Multi-Agent Simulation

Scientific goals

1. Enrich the AMAS theory with general learning mechanisms andstrengthen the coupling between micro and macro levels.

2. Propose a new generic calibration method of models.

3. Enrich GAMA tools


CALICOBA : Agent-based calibration of simulation models

IRIT – Université Toulouse 1 – Capitole


calibration, simulation, multi-agent system, AMAS    

In many fields of science and engineering, simulation is a key part in understanding phenomena or predicting their future evolution. It is also a useful tool for planners and decision-makers in order to guide them in their decisions. A simulation is a computer model of a real-world system that contains entities in interaction and that is used to understand and/or predict the evolution of the system it represents.  

 In order to be as close as possible to real phenomena, simulation models have to be calibrated. Several different calibration methods exist, from classical optimization methods to multi-agent systems and data assimilation. As most simulation models are complex systems, usual calibration methods are not really appropriate to account for dynamics changes. One way to handle those changes is to tune parameters values while the simulation is running, using data observed on the real system. One additional constraint is that models are seen as black boxes, i.e. the calibration system has no insight of the inner workings of the model. This means that the calibration system has to learn the influence of each parameter on each model observation.    

This work proposes a new online calibration method, CALICOBA, based on adaptive multi-agent systems, that aims to solve these problems. It features two kinds of agents: parameters and objectives. The role of objective agents is to estimate the distance of observations from the real data. Parameter agents take these distances to evaluate the best value the corresponding model parameter should take in order to minimize the distance computed by objective agents. Parameter agents have to learn their influence on each objective agent in order to compute the best value.    

For instance, in the case of a simple traffic model with two parameters (maximum vehicle speed and reaction time) and two observables (traffic density and mean vehicle speed). Lets assume that there is a data source for the actual roads represented by this model. The role of CALICOBA would be to take in the observed data and compare it to each observable in order to compute a distance for each. With this information, the parameter agents have to determine in which direction to modify their values in order to decrease these distances.    

As this is still a work in progress, the system has yet to be tested on traffic models, but it is planned for the near future.

Scientific goals

- Online self-calibration

- Automatic learning of model input/output interactions


Damien Vergnet –

Frédéric Amblard –

Elsy Kaddoum –

Nicolas Verstaevel –

Experimental Platform for Autonomous Vehicles

Context Presentation

Several manufacturers now offer road vehicles that are almost autonomous (Tesla, Uber, Apple, EasyMile, ...). But they still require certain situations where human control is essential, which is why research is still important in this field. Experimentation is an essential means to evaluate software in a context close to reality. Evaluation in real conditions has several disadvantages: (i) user safety is difficult to guarantee; (ii) the cost of experimentation is very high; (iii) experimental sites in urban areas are very rare; (iv) the volumes of communications with the infrastructure and the necessary processing capacities are difficult to guarantee.

In order to carry out this research and evaluate it experimentally, we are developing an environment in which model cars move in a realistic environment.

20200120_104900 - Guilhem Marcillaud 


Semi-real simulation, Intelligent Transport System

Scientific goal

•    Experiment algorithm with miniature model of vehicles


Traffic simulation at Toulouse city scale

Context Presentation

The ANR SWITCH project is focused on studying the impact of smart city new transport modes on infrastructure. Although the new transports (electric, autonomous vehicles, transport on demand, bicycles, etc.) can facilitate intra-urban mobility and improve the quality of life, they can also create new constraints for which cities must be prepared. Then urban planners need tools to evaluate the impact of urban policies in terms of mobility and infrastructure to explore “What if? " scenarios.

These new simulation tools require mobility simulations on a city scale and on different time scales. As part of the meta-model design for the Smart City simulation, this internship is about developing an urban mobility model allowing to couple a flow-based circulation model with a multi-agent model allowing fine modeling of driver behavior. The internship must carry out a proof of concept on the Gama platform of this type of system on the scale of an agglomeration.

poster - loic sadou


Traffic Simulation, GAMA, Multi-Agent Simulation, Smart City, Transport

Scientific goal

•    To validate a mesoscopic traffic multi-agent model considering various transports facilities


Automatic traffic generation with multi-agent simulation

Context Presentation

Mobility in cities is a crucial question to improve services offered to users and to decrease pollution. Traffic simulation is an interesting tool for decision-makers for city planning. Decision relevance depends on the realism of the simulation. To ensure that, we suppose that we have a map of the terrain (road topology, buildings…) and observation points that provide traffic information (throughput, direction…). They are in limited number and do not cover all the map.This internship focuses on generating car flows on the whole road network using adaptive multi-agent systems. For that, the system determines the location, speed and direction for each vehicle based on its neighborhood. The behavior of each vehicle self-adjusts to match, at the macro-level, traffic information provided by all observation points.


Figure 1: Example of traffic simulationDynameq © INRO

Scientific Goals

- Self-calibration of simulation parameters for realistic traffic flow generation

- Self-adaptation of the simulation to handle traffic dynamics


realistic traffic generation, multi-agent simulation, self-adaptation, cooperation


Simulation Energétique Dynamique d’un bâtiment en vue d’une gestion intelligente de l’énergie

Ce travail fait partie de la poursuite de l’étude d’optimisation énergétique du bâtiment ADREAM au LAAS-CNRS. L’objectif de cette thèse est l’optimisation du bilan Consommation/Production avec un focus sur la consommation électrique CVC (Chauffage, Ventilation, Climatisation). La performance énergétique du bâtiment ADREAM dépend de deux volets, un volet de production d’énergie (production d’énergie électrique par les PV, et production d’énergie thermique par son système géothermique), et un volet de consommation d’énergie qui correspond aux consommations électriques liées aux systèmes CVC, les équipements électroniques et l’éclairage. La démarche de ce travail a pour but la gestion intelligente de l’énergie électrique d’un réseau futé. Ce réseau futé est la plateforme ADREAM, qui comporte plusieurs sources et systèmes d’énergie en interaction constante. Dans cette plateforme multidisciplinaire on trouve un système géothermique liée à des PACs (Pompes à chaleur) qui servent à produire du chaud ou du froid pour le bâtiment. ADREAM comporte aussi un système de ventilation liée à un puits canadien, et une grande surface des PV (panneaux photovoltaïques) pour la production d’électricité (soit pour autoconsommation, soit pour redistribution au réseau électrique du LAAS). Ainsi, afin que le réseau puisse fonctionner d’une façon plus intelligente et efficace, plusieurs modèles précis sont développés. La calibration des modèles est réalisée selon les données existantes, récupérables par un système de supervision. Une fois qu’un modèle est calibré, des simulations sont lancées pour la prédiction de la consommation électrique en vue d’une amélioration de la régulation des systèmes.




- ALONSO Corinne (LAAS-CNRS) :


OSM To 3D : Simulation du campus de l’UPS


Certaines salles du bâtiment U4, dont le “CampusFab” sont équipées d’une multitude de capteurs permettant d’en mesurer la température, l’humidité, la luminosité, le taux de CO2 et la présence. Elles possèdent également des actionneurs qui offrent la possibilité d’agir sur les dispositifs connectés (volets, chauffage et éclairage). Les capteurs fournissent des données en temps réel qui sont centralisées et archivées au sein d’un serveur web connecté à Internet. Grâce à une API, il est possible d’accéder en temps réel aux données recueillies. Ces données peuvent être téléchargées de manière automatisée, via une grande variété de langage de programmation, de façon à être intégrées aux applications développées par les équipes en lien avec le projet neOCampus.

Objectifs scientifiques

L’objectif initial du projet OSM To 3D était de visualiser en 3D une carte issue d’OpenStreetMap. Ce sujet propose d’afficher en temps réel les données issues des capteurs disposés dans les salles du bâtiment U4. Un système d’alerte automatique permet de signaler les anomalies (haute température, forte luminosité etc...). Cela permettra d’être averti au plus vite en cas de conditions extrêmes (hautes températures, forte luminosité) et de pouvoir réagir manuellement via l’application ou de manière automatique grâce à une IA programmée pour assurer le confort des utilisateurs.


- Dorian Roques (IRIT) :

- Cédric Sanza (IRIT) :


Back to Top