(DP 1985-05) A Class of Reducible Dynamic Control Problems

Abstract

We show that analysis of control problems with a "reduction set," i.e., a subset of the control vector that satisfies certain strong conditions, avoids the "curse of dimensionality." The problem boils down to solving n static equations in n unknowns (the control plus the state variables) and leapfrogs the 2-point boundary value problem that besets more general control problems.