Nicolas Gast About me Research Teaching


PhD students

If you want to apply for a PhD thesis, the best is to start doing an internship.

Internships (for bachelor, master or PhD student)

I am happy to offer internships to students. If you want to apply for an internship, please send me the contact of at least one person who can recommend you (a faculty member of your university or a person with whom you have worked).

List of possible subjects:

Post-doc position (12 months) : Mean-field limits for small stochastic networks

Keywords : stochastic networks, mean-field limit, stochastic optimization

Stochastic models of interacting agents are used in many domains (caching systems, networks,...). The analysis of the model of n stochastic entities interacting with each others can be particularly difficult. The mean field approximation is a very effective technique to characterize the transient probability distribution or steady-state regime of such systems when the number of entities n grows very large. The idea of mean-field approximation is to replace a complex stochastic system by a simpler deterministic dynamical system. Our recent progress suggest that it is possible to extend these methods to study systems with a relatively small entities (n=10). The objective of the post-doc will be to contribute to the development of theoretical and analytical tools on this subject, for example to focus on heavy-traffic regime and heterogeneous systems.

Depending on the candidate, this project can have a more theoretical flavour or a more algorithmic one. Different applications are envisioned.

Profile : Applicants must have a PhD in Computer Science, Mathematics or Statistics. A strong knowledge in stochastic modeling or optimization is necessary. Applicants should be able to work efficiently in a research team.

Post-doc position (12 or 24 months) : Distributed Software for Control and Optimization in Smart-Grids

Keywords : software engineering, networking, distributed systems, smart grid.

Context : Smart meters are currently being deployed in many electrical distribution networks. The first motivation for the deployment of smart meters is simply to streamline the billing process; however, the smart metering IT infrastructure also offers some great potential for additional applications, far beyond mere consumption metering.

In a joint research effort, INRIA, G2Elab and the company Schneider Electric are currently developing a distributed control layer that will be deployed on top of the existing smart metering infrastructure with the aim to eliminate potential congestions in the underlying electricity distribution grid. The resources that will be controlled include flexible generators, such as small-scale photovoltaics, and flexible loads, such as electric vehicles. The rationale of the control is to increase the hosting capacity of the network for photovoltaic generators, electric vehicles and other potentially flexible electric appliances. This project aspires to become a game-changer in the way we consume electricity, with a strong positive impact on the environment and a decrease of the capital expenditures invested in the distribution grid.

Job description : The candidate will join a cross-disciplinary project currently composed of two senior researchers, one post-doc and one PhD student. The candidate will contribute to the development of a distributed software application that will be deployed on top of the smart metering IT infrastructure. The functional specifications of the application have been determined, several proofs of concepts have been successfully implemented, and the development of the distributed application is now starting. The candidate will contribute to the development of this application, with a focus on software architecture design and on the communication network layer.

The work will encompass the design, implementation, deployment and testing of a distributed control protocol. The protocol will be deployed on a distributed hardware platform. A simulator allowing to study the performance of the control algorithm under various assumptions (such as a very large number of nodes, or degraded communication capabilities) will also be developed. In addition, the post-doctoral researcher may occasionally need to analyze the control algorithm from a more theoretical viewpoint.

In particular, the candidate will work on the following issues:

  1. The capacity of the communication channels deployed for smart metering is quite limited in general; the traffic generated by the control algorithm should thus be minimized.
  2. For the same reason, the control algorithm should be made as robust as possible with respect to packet loss and network delays.
  3. The control algorithm should be capable of self-reconfiguration whenever the network topology changes.
  4. The simulator should be able to handle large test-cases efficiently. The candidate will strive to improve the computing time and memory usage of the simulator, profile its execution, and identify and solve performance bottlenecks.

Profile : The candidate should have a PhD in Software Engineering, Computer Science, or Networking / Telecommunications. The candidate should have strong programming skills (i.e., C and python). At least some basic notions of electrical engineering are required. In addition, the candidate should be strongly motivated to work on energy-related applications and show some interest for the energy transition and for the environment. Some experience as a programmer in the software industry, or as a contributor to some large and high-quality open-source project, would be appreciated. Good written and spoken communication skills in English are required. Some knowledge of French would be a plus.

Host laboratory and partners : The post-doctoral researcher will be recruited in the POLARIS team (Performance evaluation and Optimization of LARge Infrastructures and Systems) at INRIA (French Institute for Research in Computer Science and Automation). The POLARIS team activities focus on developing models and software for distributed systems such as HPC, wireless networks and smart-grid systems. The new recruit will join a project involving, in addition to members of the POLARIS team, researchers from G2Elab and from the company Schneider Electric.

Location and duration : The position will be located in Grenoble, France, in the IMAG building. The position is open initially for 12 months, with the option to renew the contract once for another 12 months.

Salary : Gross salary will be 3000 euros/months. Health insurance is included.