When considering the stability under arbitrary switching of a discrete-time linear switched system
where , the matrix belongs to a finite family
and denotes the switching law, it is known that the most stable switching laws are associated to the so-called spectrum-minimizing products of the family . The minimal rate of growth of its trajectories is called lower spectral radius and is denoted as . For a normalized family of matrices (i.e., with lower spectral radius ) that share an invariant cone (the most typical case is the positive orthant for nonnegative matrices), all the most stable trajectories starting from the
interior of lie on the boundary of a certain invariant set (called Barabanov antinorm), for which a canonical constructive procedure is available [4], that is based on the invariant polytope algorithm developed in [3]. After presenting the main results for this case, I will consider families of matrices that share an invariant multicone [1, 2] and show how to generalize some of the known results to this more general setting (Guglielmi & Zennaro, in progress [5]). These generalizations are of interest because common invariant multicones may well exist when common invariant cones do not.
References
[1] M. Brundu, M. Zennaro Multicones, duality and matrix invariance, J. Convex Anal. 26 (2019), pp. 1021–1052.
[2] M. Brundu, M. Zennaro Invariant multicones for families of matrices, Ann. Mat. Pur. Appl. 198 (2019), 571–614.
[3] N. Guglielmi, V.Yu. Protasov, Exact computation of joint spectral characteristics of linear operators, Found. Comput. Math., 13 (2013), pp. 37–97.
[4] N. Guglielmi, M. Zennaro Canonical construction of polytope Barabanov norms and antinorms for sets of matrices, SIAM J. Matrix Anal. Appl. 36 (2015), pp. 634–655.
[5] N. Guglielmi, M. Zennaro Antinorms for sets of matrices sharing an invariant multicone, in preparation (2022).