Abstract
This paper deals with an interative procedure for solving the time-optimal control problem of linear dynamical systems.
First, a switching surface of adjoint vector is defined. It is shown that the optimal control u*(t1) at the time instant t=t1 is determined by the optimal adjoint vector P*π on that surface. By using u* (t1) throughout the next Δt sec. and repeating the same procedure, the suboptimal control u(t), 0≤t≤tf, can be estimated step by step. The method to improve this technique is discussed.
Finally, several numerical examples are shown. Some of them are concerned with the off-line computation of an optimal control. Another example is concerned with a simulation of suboptimal feedback control system.
