For discrete-time linear time invariant systems with constraints on inputs
and states, we develop an algorithm to determine explicitly, the state feed
back control law which minimizes a quadratic performance criterion. We show
that the control law is piece-wise linear and continuous for both the fini
te horizon problem (model predictive control) and the usual infinite time m
easure (constrained linear quadratic regulation). Thus, the on-line control
computation reduces to the simple evaluation of an explicitly defined piec
ewise linear function. By computing the inherent underlying controller stru
cture, we also solve the equivalent of the Hamilton-Jacobi-Bellman equation
for discrete-time linear constrained systems. Control based on on-line opt
imization has long been recognized as a superior alternative for constraine
d systems, The technique proposed in this paper is attractive for a wide ra
nge of practical problems where the computational complexity of on-line opt
imization is prohibitive. It also provides an insight into the structure un
derlying optimization-based controllers. (C) 2001 Elsevier Science Ltd. All
rights reserved.