Efficient Variable Selection for LMI-Based Data-Driven Control

This paper addresses linear matrix inequality (LMI)-based data-driven control for discrete-time nonlinear systems. We propose a novel approach to decision variable selection in the LMI formulation, where the size of the decision variables is independent of the amount of available data. In contrast, existing methods involve decision variables whose size scales with the data. A numerical example demonstrates that the proposed method significantly reduces computation time compared to existing approaches.