Distributed control of the Laplacian spectral moments of a network
It is well-known that the eigenvalue spectrum of the Laplacian matrix of a network contains valuable information about the network structure and the behavior of many dynamical processes run on it. In this paper, we propose a fully decentralized algorithm that iteratively modifies the structure of a network of agents in order to control the moments of the Laplacian eigenvalue spectrum. Although the individual agents have knowledge of their local network structure only (i.e., myopic information), they are collectively able to aggregate this local information and decide on what links are most beneficial to be added or removed at each time step. Our approach relies on gossip algorithms to distributively compute the spectral moments of the Laplacian matrix, as well as ensure network connectivity in the presence of link deletions. We illustrate our approach in nontrivial computer simulations and show that a good final approximation of the spectral moments of the target Laplacian matrix is achieved for many cases of interest. © 2010 AACC.
