My primary research interests are modeling, analysis, and control of complex networked systems and engineering infrastructureswith applications in social networks, technological infrastructure, and biological systems. Since critical societal functions are increasingly dependent on complex networks, it is important to acquire a deeper understanding of the relationship between the structure and the dynamics of these systems.

Research lines of particular interest are described below:

Data-Driven Analysis and Control of Complex Networks (Support: NSF CAREER Award)

In practice, modeling and controlling real-world complex dynamical networks may be a daunting task, due to the amount of environmental uncertainties and structural volatility real networks suffer during operation. On the other hand, we are witnessing the proliferation of high-throughput devices able to sensor, monitor, and record, with unprecedented detail the structure and behavior of real-world large-scale networks. We could, therefore, utilize the massive amounts of data generated by these high-throughput technologies to overcome the issues associated with uncertain systems operating in volatile environments. Although a data-rich environment offers us the opportunity to increase the performance of complex networked systems, it also brings several engineering challenges.   

Relevant references from our research group:

Analysis of Complex Socio-Technical Systems (Support: NSF Bigdata)

The main aim of our work is to develop a theoretical framework, based on spectral graph theory and convex optimization, to compute with low computational overhead global properties of a network from local structural properties. In particular, we derive optimal bounds and estimators of spectral properties of interest from structural information. Our results are useful to unveil the set of structural properties that have the highest impact in the eigenvalue spectrum of a network. In particular, in the case of online social networks, we find that the correlation between the distribution of degrees and triangles in the network plays a key role in the spectral radius.

Relevant references from our research group:

Optimal Communication for Fast in-Network Coordination (Support: NSF NeTS)

The goal of this project is to develop a new framework where teams of mobile robots self-engineer the structure of their communication network in order to improve the performance of distributed algorithms used for team coordination. The key idea that enables this work is the representation of the network structure in terms of metrics that depend on the full eigenvalue spectrum of the network's adjacency and Laplacian matrices but can be approximated in a very efficient and decentralized way. These metrics can then be related to the performance of popular, distributed, coordination algorithms, such as consensus, gossiping, and viral information dissemination. This project closes the loop in the design of distributed coordination algorithms implemented on, generally mobile, networks by incorporating the network structure in the space of control variables that affect distributed coordination. The resulting hybrid network of mobile robots is controlled jointly in the space of network configurations and robot positions. 

Relevant references from our research group:

Optimal Resource Allocation for Epidemic Control

Understanding spreading processes in complex networks and designing control strategies to contain them are relevant problems in many different settings, such as epidemiology and public health. The aim of our research is to control spreading processes in networks by distributing protection resources throughout the nodes. In our study, we consider two types of containment resources: (iPreventive resources able to protect (or ‘immunize’) nodes against the spreading (such as vaccines in a viral infection process), and (iicorrective resources able to neutralize the spreading after it has reached a node, such as antidotes in a viral infection. In our framework, we associate a cost to these resources and study the problem of finding the cost-optimal distribution of resources throughout the network to contain the spreading. [Slides]

Relevant references from our research group:

Resilience of Critical Networked Infrastructure (Support: The Rockefeller Foundation)

Multi-agent networked systems are used to model a wide variety of systems, from robotic networks to power systems, to biological networks. The emergent dynamics of a network of dynamic agents can be strongly affected by the presence of network failures. The main of our research in this direction is to find computationally efficient methods to study the effects of link or node failures on networked infrastructure. We have provided several methods for detection and isolation of failures in a network of dynamic agents, based on the presence of discontinuities in the derivatives of the output responses of a subset of nodes.

Relevant references from our research group:

Network Identification from Dynamic Behavior

The reconstruction of networks of dynamical systems is an important task in many realms of science and engineering, including biology, physics, and finance. The aim of our research is to propose several algorithms to reconstruct the structure of a directed dynamic networks. We have proposed an algorithm to find the Boolean structure of the unknown topology based on the analysis of power spectral properties of the network response when the inputs are wide-sense stationary (WSS) processes of unknown power spectral density (PSD). Apart from recovering the Boolean structure of the network, we have also proposed an algorithm to recover the exact structure of the network (including edge weights) when an eigenvalue-eigenvector pair of the connectivity matrix is known.

Relevant references from our research group: