Robust H∞-optimal Output Feedback Actuator Placement with a Class of Actuator Constraints

This paper investigates the actuator placement of uncertain linear systems. The designed actuator locations are subject to a class of nonlinear equality constraints, and a dynamic output feedback controller should also be derived based on measured outputs. To minimize H infinity norm of the closed-loop systems, the actuator placement problem is formulated as an uncertain non-convex optimization problem. A series of transformations are developed to convert this problem into a deterministic successive convex problem. Correspondingly, an optimization algorithm is derived based on prediction-correction procedure.