Control of convex-monotone systems
Anders Rantzer, Lund University, Sweden
Abstract:
We define the notion of convex-monotone system and prove that for such systems the state trajectory is a convex function of the initial state and the input trajectory. This observation gives a useful class of nonlinear dynamical systems for which optimal trajectories can be performed by convex optimization. Applications to evolutionary dynamics of diseases and voltage stability in power networks are presented. In particular, first order convex optimization methods enable computation of optimal trajectories with a complexity that grows only linearly with the number of network edges.