Jan Maciejowski

Professor Maciejowski’s main interest is in the control of dynamic systems using real-time optimisation. The framework used is model predictive control (MPC). Algorithmic developments are being pursued such as ‘Multiplexed MPC’ for fast computation with multi-input systems, and ‘Lasso-MPC’ for generating sparse controls when there are redundant actuators or when control is expensive. The combination of MPC with machine-learning methods is also being investigated, with a possible application being fault-tolerant control. The solution of MPC problems on special-purpose hardware such as FPGAs and GPUs is an area of current interest. Applications of MPC to problems of spacecraft control, flight control, air-traffic management, and energy reduction in paper-making have been studied in recent years. Application to industrial networks containing both generators and consumers of electrical and thermal power is a new area of investigation.