This paper considers a kind of robust pole assignment problem for discrete-
time systems. A recently developed rr posteriori measure of closed-loop eig
envalue sensitivity suitable For MIMO deadbeat regulator design is minimize
d during the search of the state feedback gain. The problem is cast into an
LMI-based convex optimization task in which constraints on the size of fee
dback gain can be catered naturally. The proposed pole assignment algorithm
can also be used to find the minimum feedback gain with respect to differe
nt norms. A numerical example is then employed to illustrate the effectiven
ess of the method. Copyright (C) 2000 John Wiley & Sons, Ltd.