Abstract:Network collapse, characterized by the abrupt transition between distinct states within a complex networked system, carries significant implications encompassing catastrophic repercussions and substantial societal costs, including power grid failures and disease outbreaks. Despite notable advancements, the intricate interplay between the hierarchical structure and dynamics of networks in shaping the collapse process remains inadequately understood. In this study, we establish a mathematical framework that facilitates the reduction of the dimension of any N -dimensional dynamical network to a kmax shell-dimensional rendition by imposing the k-core as a governing constraint upon the network's structure and dynamics. Subsequently, we employ this framework to elucidate the collapse process of the dynamical network. Our investigation underscores that the patterns of network collapse are intrinsically linked to the influence exerted by the network hierarchy upon the collapse process. Notably, dynamic networks may undergo either hierarchical or simultaneous collapse contingent upon the hierarchical impact on the network's collapse progression. In scenarios where hierarchy holds sway, the network attains the tipping point of collapse upon the collapse of nodes within the maximum k-core. Conversely, when the hierarchy's influence is absent, nodes within the dynamic network succumb to simultaneous collapse. Furthermore, we introduce a robust criterion for prognosticating the tipping points of dynamic network collapse. Our explorations reveal that it suffices to consider nodes and links within the maximum k-shell and t-shell with βtt<βtk in the network, provided the collapse process is entwined with network hierarchy, to anticipate the tipping point.

Dimension Reduction of Collective Attention Networks.

Empirical Analysis of Attention Behaviors in Online Social Networks

Dimensionality reduction of networked systems with separable coupling-dynamics: theory and applications

Dimensionality Reduction in Discrete-Time Dynamical Systems

Dimension reduction of dynamics on modular and heterogeneous directed networks

Dimensionality Reduction in Stochastic Complex Dynamical Networks

Generalized dimension reduction approach for heterogeneous networked systems with time-delay

Predicting Tipping Points in Mutualistic Networks Through Dimension Reduction

Attention Dynamics in Collaborative Knowledge Creation

Multiscale dynamical embeddings of complex networks

Dimension reduction in higher-order contagious phenomena

An Attentional Multi-scale Co-evolving Model for Dynamic Link Prediction

An individual-based model of collective attention

Exploring Collective Dynamics in Communication Networks

Time-Dependent Collective Behavior in A Computer Network Model

Dimensionality Reduction of Collective Motion by Principal Manifolds

An Attention Model for the Formation of Collectives in Real-World Domains

Collapse process prediction of mutualistic dynamical networks with k-core and dimension reduction method

Recovering Network Structures Based on Evolutionary Game Dynamics Via Secure Dimensional Reduction.

A coevolving model based on preferential triadic closure for social media networks

A New Collaborative Neural Network: Using HDMR for Dynamic Structural Adaptation*