A fast algorithm for the computation of an upper bound on the \(\mu\)-norm (Q1975581)

The authors propose a fast numerical algorithm for the computation of an upper bound for the structured singular-value norm. The algorithm is based on the equivalence between the singular value problem and the Hamiltonian matrix eigenvalue problem. More precisely, using an iteration scheme, sets of frequency intervals are computed where maximizers for the structured singular-value norms are known to lie. The rate of convergence for the algorithm is found to be quadratic. The results of numerical experiments are presented and demonstrate great computational savings in comparison with a grid search. The algorithm may be extended to compute an upper bound on the mixed structured singular value norm which could be applied to systems with mixed dynamic and real parametric uncertainty.