Thematic Semester: Stochastic control and learning for complex networks

18 Mars

Thematic Semester: Stochastic control and learning for complex networks

Date: June to December 2024



Stochastic networks is both the name and the domain of study of a wide variety of complex large-scale systems, such as computer, transport, and electrical networks, in which components interact randomly subject to a structure described by a graph. Their analysis often requires developing and combining concepts from graph theory, to describe interactions between agents, and stochastic modeling, to understand how these interactions impact the long-term dynamics. While many classical studies have focused on systems in which either the graph is highly symmetrical, or components interact only within a local neighborhood of the graph, more recent studies have tackled systems, inspired by modern applications like data centers, involving more sophisticated interactions. This trend is fueled by the growing popularity of machine learning, which has initiated a paradigm shift in the stochastic network community that is visible in top-tier publication venues.

In this context, the goals of the semester are (i) to understand the challenges and opportunities in the application of machine learning to the analysis and optimization of stochastic networks, and (ii) to develop a reflection on the future of stochastic networks as a field of study.

Concretely, we will organize the following workshops on methodological aspects and applications of machine learning for stochastic networks:

Organized by SOLACE CIMI Team
Urtzi Ayesta,, IRIT–CNRS
Olivier Brun,, LAAS–CNRS
Céline Comte,, LAAS–CNRS
Matthieu Jonckheere,, LAAS–CNRS
Balakrishna Prabhu,, LAAS–CNRS
Ina Maria Verloop,, IRIT–CNRS

Sponsored by:

Long-term visitors
- Vivek Borkar (Indian Institute of Technology Bombay)
- Itai Gurvich (Northwestern University)
- Sean Meyn (University of Florida)
- Adam Wierman (Caltech)