U-centrality: A Network Centrality Measure Based on Minimum Energy Control for Laplacian Dynamics

A team of researchers from the University of California, San Diego, and other institutions has published a paper titled 'U-centrality: A Network Centrality Measure Based on Minimum Energy Control for Laplacian Dynamics' on arXiv. The work addresses a fundamental limitation in network science: traditional centrality measures like degree, betweenness, and eigenvector centrality are purely structural. They quantify a node's importance based on its position in the network graph but often ignore the actual dynamics—like information flow, opinion diffusion, or synchronization—that occur over that structure. The authors argue that a node's importance is inherently context-dependent and must reflect both the system's dynamics and the specific operational objectives, such as efficiently unifying opinions in a social network or controlling a power grid.

To bridge this gap, the researchers propose a dynamic, task-aware centrality framework rooted in optimal control theory. They formulate a problem of achieving a unified average opinion across a network with minimal control energy, modeled by Laplacian dynamics—a common framework for diffusion and consensus processes. By analyzing the variance of the system's terminal state, they derive a novel metric called U-centrality, which quantifies a node's 'unifying' power. A key theoretical finding is that U-centrality interpolates between known measures: for very short control time horizons, it aligns with simple degree centrality (a node with many connections is easiest to control quickly). Over longer time scales, it converges to a new centrality closely related to current-flow closeness centrality, which considers all possible paths for influence.

This work, presented at the 2025 IEEE Conference on Decision and Control (CDC), provides a mathematically rigorous tool that connects structural graph theory with dynamical systems and control theory. It moves beyond static snapshots of networks to a more holistic view where a node's importance is defined by how effectively it can steer a dynamic process toward a desired goal with minimal effort. The framework is versatile and can be adapted to various networked systems where control and influence are key, from social media and epidemiology to robotics and infrastructure management.