Skip to Main content Skip to Navigation

Comprendre le fonctionnement de simulations sociales individus-centrées: application à des modèles de dynamiques d'opinions

Abstract : Using an individual-based approach, particularly for the modelling of social systems, the modeller often faces a major source of complexity within his model. The relation between individual behaviours and collective ones is a source of complexity present in the real world that is reproduced within individual-based models. Then he faces a model which functioning is difficult to understand even if he knows the rules of functioning at the individual level. We propose here the use of an experimental approach to understand individual-based social simulation models. With the aim to encircle the subject of this approach, the model, we first expose a formal framework to express individual-based models. This framework will enable us to include several types of individual-based models, namely cellular automata, microsimulation or multi-agent simulation. We will then present individual-based models of social influence of the literature, using this framework as a guideline, with the aim to represent the broad range of possible behaviours we can introduce within the model. In a second part, we will focus to the experimental approach as a mean to understand the behaviour of model. We will first present the different viewpoints on the individual-based model that the modeller can adopt (individual, collective, simulations population among others) and we will argue about the necessary complementarity between those viewpoints with the aim to understand the functioning of the model. We will propose thereafter a software environment dedicated to the management of experimental designs on simulation models. This latter will help the systematic exploration of the parameters space of a model. We will present then an incremental methodology for building models to enable us to obtain a progressive understanding of the functioning of complex models seen as a collection of models of increasing complexity. In the last part, we will apply the exposed elements presenting a set of models of opinion dynamics whose understanding is enrich by studying first very simple models and then complexifying progressively the studied models. We will then present successively a model of conditional interaction with a homogeneous population. We will make evolve the interaction process in a second model and then we will explore this latter introducing heterogeneity in the population and a social network constraining the interactions.
Document type :
Complete list of metadatas

Cited literature [185 references]  Display  Hide  Download
Contributor : Frédéric Amblard <>
Submitted on : Monday, October 6, 2008 - 11:22:15 AM
Last modification on : Thursday, December 17, 2020 - 10:50:06 AM
Long-term archiving on: : Friday, June 4, 2010 - 12:14:50 PM


  • HAL Id : tel-00326863, version 1



Frédéric Amblard. Comprendre le fonctionnement de simulations sociales individus-centrées: application à des modèles de dynamiques d'opinions. Autre [cs.OH]. Université Blaise Pascal - Clermont-Ferrand II, 2003. Français. ⟨tel-00326863⟩



Record views


Files downloads