A technique different from the boundary layer method is developed to deal with singularly perturbed optimal control problems. The technique is applicable, in particular, in the case when the optimal control takes the form of fast oscillations and the boundary layer method cannot be used. Necessary and sufficient conditions for the optimal control to be slow are given. The results are illustrated by examples.
|Number of pages||22|
|Journal||SIAM Journal on Control and Optimization|
|Publication status||Published - Sep 1992|