Design of critical control systems using genetic algorithms and LMI-based regional pole assignment technique

A methodology for designing critical systems subject to magnitude- and rate-limited reference inputs is proposed. The worst-case tracking error is minimized under the constraint on the control effort. The LMI-based technique developed by Chilali and Gahinet is utilized to assign the resulting closed-loop poles into the intersection of vertical strips and a conic sector. The optimal pole region is searched by means of the simple genetic algorithm. An illustrative design example of the position control system for a two-inertia system is provided to demonstrate the proposed design methodology.