My research activity is part of the SMAC team of the I2C thematic group whose head is Pr. Philippe Mathieu. Within the CRIStAL laboratory, my work concerns the coding of artificial intelligence behaviors, called agents, in interaction. By concretizing these interactions locales and minimizing the number of mechanisms/parameters, my goal is to bring out complex processes that cannot be modeled either by predictive equations and neither by solvable equations. The applications include the development of methods for distributed problem solving and explanatory models for social simulation.

My scientific project, which is based on international and interdisciplinary cooperation, is split in 2 axes.

The first axis consists in developing the techniques of distributed artificial intelligence for big data processing. I study the problem of reallocation of tasks for balancing MapReduce jobs. I propose distributed strategies, which are dynamic and adaptive to the data, the computing environment and the runtime hazards. The scientific lock consists in design a resource allocation process that results from the asynchronous execution of individual behaviors. These works are carried out in collaboration with Pr. Kostas Stathis from RHUL (Royal Holloway University of London).

The second axis consists in proposing an explanatory model of social phenomena. The individual-centered approach makes it possible to understand the macroscopic and mesoscopic processes that result from individual behaviors in online communities. The scientific lock consists in calibrating the model to from real and large data. These works are carried out in collaboration with Pr. Fabien Éloire at Clersé (Lillois Center for Studies and sociological and economic research) in the University of Lille.