Reviving networked multi-dimensional dynamical systems

TL;DR

This paper introduces a novel control method for multi-dimensional dynamical systems, enabling the activation of network connections by controlling a single node, which simplifies complex systems for effective state restoration.

Contribution

It presents a new approach to control multi-dimensional systems by dimensionality reduction based on shortest paths, extending control methods beyond one-dimensional systems.

Findings

Successfully controls real multi-dimensional systems

Reduces complex systems to simple models for analysis

Predicts control effort needed for system restoration

Abstract

From gene regulatory networks to mutualistic networks, controlling a single node in the network topology can transform these complex dynamical systems from undesirable states to desirable ones. Corresponding methods have been well-studied in one-dimensional dynamical systems. However, many practical dynamical systems require description by multi-dimensional dynamical systems, such as the mutualistic symbiotic systems formed by flowering plants and pollinating insects. Existing one-dimensional methods cannot handle the cases of multi-dimensional dynamical systems. Based on this, we propose a method to control a single node to activate network connections in multi-dimensional dynamical systems. In such systems, the changes of each node are described by multiple nonlinear differential equations. All remaining nodes are stratified according to the shortest path to the controlled node, thereby reducing the dimensionality of the system. Such a large-scale dynamical system can ultimately be replaced by a very simple system. By analyzing the reduced-dimensional system, we can predict the extent of control needed to restore the system state. We apply this method to a wide range of fields, achieving activation of various real multidimensional complex dynamical systems.