An observational equivalence among H-infinity-control policies

This paper constructs an equivalence class of objective functions that prod uce the same linear state-feedback decision rule. Members of this class are distinguished by the relative weight placed on state variability and the m inimized H-infinity norm of the system's transfer function. This result is used to show that in models of dynamic policy formation a concern for robus tness in the presence of model uncertainty can substitute for the ad hoc in corporation of adjustment costs. (C) 1999 Elsevier Science S.A. All rights reserved.