Optimal stability and eigenvalue multiplicity (Q5943879)

The authors consider the problem of choosing a linearly parametrized square matrix so that its eigenvalues are as far into the left half-plane as possible. Such problems are prototypical in robust control and stability analysis. Generalizing their earlier work [Proc. Am. Math. Soc. 129, No. 6, 1635-1642 (2001; Zbl 0965.15020)], they show that the optimal choice of parameters typically corresponds to patterns of multiple eigenvalues. Under certain nondegeneracy conditions (including requiring the critical matrix to have nonderogatory spectral abscissa) they show that the multiplicities of the active eigenvalues at a critical matrix are unchanged under small perturbations.