In this paper, the problem of determining the worst-case Hz performanc
e of a control system subject to linear time-invariant uncertainties i
s considered. A set of upper bounds on the performance is derived, bas
ed on the theory of stability multipliers and the solution of an origi
nal optimal control problem. The numerical issues raised by the result
ing computational problems are discussed, in particular, newly develop
ed interior-point convex optimization methods, combined with linear ma
trix inequalities, apply very well to the fast and accurate solution o
f these problems. The new results compare favorably with prior ones. T
he method can ha extended to other types of perturbations.