Performance Equivalence of Suboptimally Controlled Nonlinear Systems
by R. C. Durbeck
A procedure is described that shows how a technique used to develop performance bounds for a large class of nonlinear dynamic systems with state-dependent control policies can be extended to determine whether a nonlinear system can be controlled so that it is at least "performance equivalent" to an associated optimally controlled linear system. A procedure for generating one or more control policies to attain this equivalence is also discussed. An example illustrates the fact that more than one control policy may satisfy the equivalence criterion.
PDF