In the area of personalized and precision medicine, one of the main goals is the design of decision support software for diagnosis and therapy. In this perspective, the main challenge is the lack of methods to interpret the data of the “omic”. Focusing on the prediction of therapeutic targets based on the analysis of molecular networks, we formalized adapted theoretical frameworks in order to infer the causal targets responsible for phenotypic transition (healthy / sick). Our methodology is implemented in a set of software tools based on game theory and abduction principles applied to Boolean networks.

We also worked on the modeling, simulation and formal verification of decision modules of autonomous and communicating vehicles. We have proposed in this framework an original, scalable and parametric modeling in timed automata and also a modeling of multi-agent simulation, both of which make it possible to evaluate the aspects of security, fluidity and comfort of various decision policies. We have shown in particular, by developing a comparison between these two approaches, what impact the level of abstraction chosen on the indicators considered had. Our simulation method has been implemented in a plugin for the Gama simulator.

We are also interested in Cloud Computing and IoT (Internet of Things) technologies in the context of the Internet of the Future. The deployment of these technologies involves many problems of great complexity. We then proposed several approaches to the modeling of these problems in order to relax some constraints by a better understanding of the specificities of these technologies and their use. We also addressed the problem of holistic and autonomic control of a converging “Cloud IoT” environment meeting the time and processing resource requirements for collecting data from thousands of potentially deployed sensors in the environment. Finally, we proposed a modeling based on Petri nets autonomic mechanisms to manage the elasticity of resources in the Cloud, whose analysis was conducted using our SNAKES tool.

Other applications were studied. In particular we proposed a formal framework for modeling and analysis of distributed storage systems, for example cache hierarchies in high performance computing architectures in collaboration with the CEA. More recently, a collaboration with INRA Montpellier has started on ecosystem modeling and analysis.