Master Stability Functions in Complex Networks

TL;DR

The paper presents the Master Stability Function (MSF) as a unifying framework to assess synchronization stability in complex networks. It derives and extends the MSF for identical systems with diffusive and natural coupling, directed graphs, multilayer structures, and higher-order interactions, illustrating principles with a Rössler-oscillator example. By decoupling network topology (Laplacian spectrum) from node dynamics, MSF provides a practical stability criterion via $\Lambda^{\max}(\kappa) < 0$ across diverse settings, with explicit conditions linked to Laplacian eigenvalues and coupling parameters. It also outlines algorithmic approaches, discusses classifications of MSF curves, and points to future directions, including time-series methods and machine-learning integration, to broaden applicability in real-world systems.

Abstract

Synchronization is an emergent and fundamental phenomenon in nature and engineered systems. Understanding the stability of a synchronized phenomenon is crucial for ensuring functionality in various complex systems. The stability of the synchronization phenomenon is extensively studied using the Master Stability Function (MSF). This powerful and elegant tool plays a pivotal role in determining the stability of synchronization states, providing deep insights into synchronization in coupled systems. Although MSF analysis has been used for 25 years to study the stability of synchronization states, a systematic investigation of MSF across various networked systems remains missing from the literature. In this article, we present a simplified and unified MSF analysis for diverse undirected and directed networked systems. We begin with the analytical MSF framework for pairwise-coupled identical systems with diffusive and natural coupling schemes and extend our analysis to directed networks and multilayer networks, considering both intra-layer and inter-layer interactions. Furthermore, we revisit the MSF framework to incorporate higher-order interactions alongside pairwise interactions. To enhance understanding, we also provide a numerical analysis of synchronization in coupled Rössler systems under pairwise diffusive coupling and propose algorithms for determining the MSF, identifying stability regimes, and classifying MSF functions. Overall, the primary goal of this review is to present a systematic study of MSF in coupled dynamical networks in a clear and structured manner, making this powerful tool more accessible. Furthermore, we highlight cases where the study of synchronization states using MSF remains underexplored. Additionally, we discuss recent research focusing on MSF analysis using time series data and machine learning approaches.

Figures (5)

• Figure 1: Portray the master stability function (MSF) in different networks and dynamical systems. To predict the stability of synchronization, we analyze the MSF, which depends on the generic parameter $\kappa$. The parameter $\kappa$ (Eq. (\ref{['mse1']})) encapsulates the influence of network topology through the Laplacian matrices. The functional form and values of MSF are determined by the dynamics of coupled oscillators, such as the Lorenz, Rössler, and other nonlinear dynamical systems (Table \ref{['table_MSF']}).
• Figure 2: Different classes of MSF. The MSF $\Lambda^{max}$ is plotted as a function of $\kappa$. The MSF can be classified into different classes based on the number of crossing points with the $x$-axis. For example, here we display five classes, $\mathcal{C}_0$ (red curve), $\mathcal{C}_1$ (purple curve), $\mathcal{C}_2$ (green curve), $\mathcal{C}_3$ (yellow curve) and $\mathcal{C}_4$ (brick red curve), and these classes have crossing points zero, one, two, three and four with the $x$-axis respectively.
• Figure 3: Master Stability Analysis of coupled Rössler Oscillators. To analyze master stability, we take (a) a network of $N=10$ nodes with random configuration. (b) Identical Rössler oscillators with parameters $a = 0.15, b = 0.2, \text{ and } c = 10.0$ are placed on the nodes of this network, and these oscillators are coupled in the $x$ component. To find the range of the coupling parameter for stable synchronization, (d) we can determine the average synchronization error from Eq. (\ref{['sync_err']}) or (e) the Lyapunov exponents. From (d) and (e), we can observe that the synchronization state is stable in the range $(\sigma_1\leq \sigma\leq\sigma_2)$ where $\sigma_1\approx 0.26$ and $\sigma_2\approx 6.0$. The same $\sigma$ range can be retrieved from the Master Stability Analysis as in (c). (c) The minimum and maximum nonzero eigenvalues of the Laplacian matrix of the network in (a) are $\mu_2 = 0.77, \text{ and } \mu_N = 6.89$, respectively. Further using the values of $\kappa_1$ and $\kappa_2$ from (f) we can determine the range of the coupling strength $\sigma$ where the synchronization state of the coupled Rössler oscillators is stable. (f) The Master Stability Function for $x$ component coupled Rössler oscillator $(\Lambda^{max})$ is plotted as a function of generic coupling function $\kappa$. The MSF is negative in the range $\kappa_1=0.2$ and $\kappa_2=6.0$. This defines the stable region $R_{\kappa}=\{(\kappa_1,\kappa_2) \vert \Lambda^{max} < 0\}$.
• Figure 4: Master Stability Function of Rössler oscillators underlying various coupling schemes. We implement coupling in different variables and plot the MSF as a function of the generic parameter $\kappa$. Here the coupling $x \rightarrow x$, $y \rightarrow x$ and $z \rightarrow x$ means the $x$ variable of the $i$th oscillator is coupled through the $x$, $y$ and $z$ variable of the $j$th oscillator respectively and others follow. Based on the transitions from unsynchronized state to synchronized state and vice-versa, we can classify the different classes of the MSF as discussed in section \ref{['msf-class']}.
• Figure 5: Schematic representation of the two types of multilayer networks. (a) A multilayer network without interlayer connections. For example, within the same layer, nodes can have two types of independent connections, such as two modes of transportation within a city, denoted by solid green and dashed red lines and represented by Laplacians $\bm{L}^{(1)}$ and $\bm{L}^{(2)}$. (b) A two-layer multilayer network includes both intra-layer and inter-layer connections. An example of this is social interactions across different platforms, such as LinkedIn and Twitter. Here, intra-layer interactions within each layer are represented by solid lines and encoded in the Laplacian matrices $\bm{L}^{(1)}$ and $\bm{L}^{(2)}$. The inter-layer interactions, which connect copies of the same node across layers, are represented by dashed lines and encoded in the Laplacian matrix $\mathcal{B}$ (Eq. (\ref{['intralayer_lsup_lap_mlil']})).