The Cooperative Multi-Agent System (SMAC) group of the IRIT Lab in Toulouse, France is offering 2 PhD positions:

both positions are related to the ANR Project GEN* (http://www.genstar.fr) which aims at providing tools and methods for the generation on synthetic populations for agent-based social simulations.
The PhDs will conduct their research at the University Toulouse 1 Capitole (http://www.ut-capitole.fr) in the IRIT laboratory (http://www.irit.fr).

Position 1: Models for the generation of artificial social networks for multi-agent simulation

In the frame of the ANR project Genstar (http://www.genstar.fr) which funds the PhD position, we are led to propose tools that allow, upstream of social multi-agent simulations (eg on epidemiological models, models dynamic opinions, …), to generate synthetic populations of individuals located and connected to each other via a social network based on demographic statistics usually available.

At present, the available models to generate artificial social networks (Small-World, Preferential Attachment, …) are usually too abstract and do not capture accurately the properties of real social networks. In contrast, ad hoc models are often used to build social networks on specific simulations taking into account the particularities of the modeled phenomena that make them difficultly transferable to other cases. The objective of this thesis is to lie halfway between abstract models and ad hoc solutions and propose models of social networks that have similar properties to empirical social networks, as well as generic methods of generation. The proposed models will be evaluated in the light of empirical evidence and applied to specific cases studies on the project (spread of dengue in India, mobility on intra-urban French cities, etc …).
The candidate must have a classical training in IT, with a good level in graph theory and social network analysis, a good knowledge of statistics, multi-agent simulation is a plus. In particular, the practice of a platform for multi-agent simulation (Gama, Netlogo …) and statistical software (R) is recommended.

Position 2: Perturbation and self-adaptation of social networks

In the frame of the ANR project Genstar (http://www.genstar.fr) which funds the PhD position, we are led to propose tools that allow, upstream of social multi-agent simulations (eg on epidemiological models, models dynamic opinions, …), to generate synthetic populations of individuals located and connected to each other via a social network based on demographic statistics usually available.

The thesis will build on existing work on the generation of social networks (Small World, Preferential Attachment, …) and will focus on the perturbation of generated social networks:

• Modification of the underlying population (adding or removing nodes in the social network, models of network growth/decay)
• Modification of the social structure (add / remove links in the social network, structure of subpopulations in communities …)
• Endogenous or exogenous perturbation of social networks (eg an epidemic diffusion leading to a change in the social network)

The candidate must have a classical training in IT, with a good level in graph theory and social network analysis, a good knowledge of statistics, multi-agent simulation is a plus. In particular, the practice of a platform for multi-agent simulation (Gama, Netlogo …) and statistical software (R) is recommended.

For content issues you may contact: Frédéric Amblard and Benoit Gaudou
Applications should be sent no later than September 15, 2014 and should include a CV, and letter of application.