Efficient algorithms for constrained minimization of infinite horizon
predictive control costs are presented. Control laws with guaranteed s
tability and asymptotic hacking are derived on the basis of appropriat
ely chosen one-dimensional and ellipsoidal constraint set approximatio
ns. These achieve comparable performance with existing QP-based contro
l laws with significant reductions in computational burden. (C) 1998 P
ublished by Elsevier Science B.V. All rights reserved.