Paper detail

Fundamental Limits and Tradeoffs on Disturbance Propagation in Large-Scale Dynamical Networks

We consider performance deterioration of interconnected linear dynamical networks subject to exogenous stochastic disturbances. The focus of this paper is on first-order and second-order linear consensus networks. We employ the expected value of the steady state dispersion of the state of the entire network as a performance measure and develop a graph-theoretic methodology to relate structural specifications of the underlying graphs of the network to the performance measure. We explicitly quantify several inherent fundamental limits on the best achievable levels of performance in linear consensus networks and show that these limits of performance are merely imposed by the specific structure of the underlying graphs. Furthermore, we discover new connections between notions of sparsity and the performance measure. Particularly, we characterize several fundamental tradeoffs that reveal interplay between the performance measure and various sparsity measures of a linear consensus network. At the end, we apply our results to two real-world dynamical networks and provide energy interpretations for the proposed performance measures. It is shown that the total power loss in synchronous power networks and total kinetic energy of a network of autonomous vehicles in a formation are viable performance measure for these networks and fundamental limits on these measures quantify the best achievable levels of energy-efficiency in these dynamical networks.