In this note, we compare the value of nonlinear and linear control for
a multiobjective H-infinity control problem with a linear time-invari
ant (LTI) plant. We show that if the H-infinity performance of a nonli
near feedback controller is measured by the maximum incremental gain,
the optimal achievable performance with nonlinear, time-varying contro
l is identical to that achievable by LTI control. (C) 1997 Elsevier Sc
ience B.V.