Controllability of Large-Scale Networks: The Control Energy Exponents
Abstract
In this article, we characterize the control energy behavior of large-scale linear network systems controlled by a single input. Specifically, we establish conditions under which two widely used control energy metrics, namely: 1) the inverse of the minimum eigenvalue and 2) the normalized determinant of the controllability Gramian, grow exponentially fast in the system dimension and derive closed-form expressions for their asymptotic exponential growth rate. We apply our findings to structured and random network architectures.