SIDDARTA is an innovative research project focused on sustainability in data-driven control of complex systems. It aims to balance resource frugality and performance in large-scale, interconnected systems by leveraging novel control paradigms, statistical physics, and machine learning. Our goal is to create more efficient, adaptable, and sustainable control methodologies.
SIDDARTA will introduce a new theoretical approach for data-driven control of complex interconnected systems, aiming at parsimony and sustainability, which are crucial challenges in many modern technological applications. Our new algorithms will address the massive paradigm shift towards data-driven control of complex systems, and propose sustainable alternatives to the current trend of evermore resource-voracious methods.
This work package focuses on developing control techniques that prioritize computational efficiency. By allowing approximate control laws and leveraging online optimization, we aim to achieve an optimal tradeoff between control performance and computational demand.
WP2 aims to create robust, model-free control strategies for highly complex systems. This involves designing adaptive sampling techniques and leveraging scenario-based optimization to efficiently manage uncertainties.
This package focuses on learning the underlying structure of complex systems and designing control strategies that optimize resource usage. By leveraging network theory and statistical physics, we aim to create multi-layered control frameworks.
WP4 translates theoretical findings into practical applications by developing a control toolbox and benchmarking its effectiveness using real-world scenarios, including multi-agent robotic systems and transportation networks.
April 24, 2025
Annual meeting of the FNRS Doctoral School on Complex Systems
For any information, please contact the PI of the project, Raphael Jungers.