Multiobjective robust control using evolutionary algorithms

An evolutionary algorithm approach is proposed for the robust design of control systems. The evolutionary algorithm is used in conjunction with an H/sub /spl infin// loop-shaping design procedure to perform multiobjective search over a set of possible weighting function structures and parameter values in order to satisfy a number of conflicting design criteria. An example of an electromagnetic suspension control system for a critical system, a magnetically levitated passenger transport vehicle, is used to demonstrate the proposed approach.