Scalable fractional-order MPC and multi-agent solvers

Develop computationally scalable algorithms for model predictive control and multi-agent control/game solvers for fractional-order systems that manage history dependence efficiently while providing performance and stability guarantees.

Background

The paper details the ‘curse of history’ in fractional prediction and surveys mitigation strategies (windowing, sum-of-exponentials compression, diffusive augmentation, and CRONE approximations).

Despite these tools, achieving scalability with certified performance and stability across single- and multi-agent settings is identified as an open challenge.