GENERALIZED ROBUSTNESS OF OPTIMALITY OF LINEAR QUADRATIC REGULATORS

Robustness of the optimality of linear quadratic (LQ) optimal regulators is studied. It is shown that under a mild assumption, optimality of LQ regulators is preserved under small perturbations on both F and G matrices. This is made possible by allowing modifications on both weighting parameters Q and R, and it also extends the existing results where perturbations on F only are allowed. It is also observed that this modification leads to a possible change in gain margin. A numerical example is discussed in detail to illustrate the results. © 1990 Taylor and Francis Group, LLC.