Current browse context:
nlin
Change to browse by:
References & Citations
Nonlinear Sciences > Adaptation and Self-Organizing Systems
Title: Controlling Statistical Moments of Stochastic Dynamical Networks
(Submitted on 4 Dec 2015 (v1), last revised 20 Jun 2016 (this version, v4))
Abstract: We consider a general class of stochastic networks and ask which network nodes need to be controlled, and how, to stabilize and switch between desired metastable (target) states in terms of the first and second statistical moments of the system. We first show that it is sufficient to directly interfere with a subset of nodes which can be identified using information about the graph of the network only. Then, we develop a suitable method for feedback control which acts on that subset of nodes and preserves the covariance structure of the desired target state. Finally, we demonstrate our theoretical results using a stochastic Hopfield network and a global brain model. Our results are applicable to a variety of (model) networks, and further our understanding of the relationship between network structure and collective dynamics for the benefit of effective control.
Submission history
From: Dmytro Bielievtsov [view email][v1] Fri, 4 Dec 2015 16:11:30 GMT (1072kb,D)
[v2] Mon, 21 Dec 2015 21:21:08 GMT (1072kb,D)
[v3] Fri, 17 Jun 2016 14:37:26 GMT (3048kb)
[v4] Mon, 20 Jun 2016 00:33:16 GMT (3048kb)
Link back to: arXiv, form interface, contact.