On Yukawa Potential Centrality for Identification of Influential Spreaders in Complex Networks

TL;DR

This paper introduces Yukawa Potential Centrality (YPC), a novel, scalable network influence measure inspired by physics, which improves identification of influential spreaders by adapting to local network structures.

Contribution

The study proposes YPC, a new non-interactive, physics-inspired centrality measure that reduces computational complexity and better captures influence in complex networks.

Findings

YPC correlates strongly with epidemic spreading models.

YPC effectively identifies key spreaders in irregular networks.

YPC is computationally more efficient than traditional models.

Abstract

Identifying influential nodes in complex networks is a fundamental challenge for understanding how information, influence, and contagion propagate through interconnected systems. Conventional centrality measures, particularly gravity-based models, often depend on pairwise interaction forces and a fixed radius of influence, which oversimplify the heterogeneous and dynamic nature of real networks. To overcome these limitations, this study proposes a novel non-interactive, action-based model, termed Yukawa Potential Centrality (YPC), which adapts the physical Yukawa potential to the topology of complex networks. Unlike gravity models, YPC computes a scalar potential for each node rather than pairwise forces, dynamically adjusting its radius of influence according to local structural properties. This formulation establishes a physically interpretable bridge between potential theory and network science, while significantly reducing computational complexity, from quadratic to near-linear time. The model is evaluated across both synthetic and real-world social networks, and its node rankings are compared with classical centrality indices and epidemic spreading models (SI and SIS). Experimental findings reveal that YPC exhibits a strong positive correlation with the SIS model and effectively isolates key spreaders, even within highly irregular topologies. These results demonstrate that YPC provides a scalable, adaptive, and theoretically grounded framework for influence analysis in social, biological, and communication networks.