Global synchronization of multi-agent systems with nonlinear interactions
Anthony Couthures, Vineeth S. Varma, Samson Lasaulce, Irinel-Constantin Morărescu · 2025
Read Global synchronization of multi-agent systems with nonlinear interactions by Anthony Couthures, Vineeth S. Varma, Samson Lasaulce, Irinel-Constantin Morărescu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-10205.txt.
Introduction
Couthures, Varma, Lasaulce, and Morarescu study synchronization in multi agent systems in twenty twenty-five. Agents adjust their states through signals from each other over time. Those signals can be nonlinear and can include bias, quantization, or state dependent error. The team considers a broad class of monotone signal functions.
Their analysis shows that synchronization states are exactly the fixed points of the signal function. They derive simple stability rules based on whether the signal under or over estimates agent states near those points. Network structure plays a key role in whether synchronization is reached in the long run. The results link communication nonlinearity with network connectivity. The authors say the work can guide coordination strategies in complex systems.
Catalog record
| CREATOR | Anthony Couthures, Vineeth S. Varma, Samson Lasaulce, Irinel-Constantin Morărescu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 34 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-10205
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Global synchronization of multi-agent systems with nonlinear interactions Anthony Couthures, Vineeth S. Varma, Samson Lasaulce, Irinel-Constantin Morărescu Abstract The paper addresses the synchronization of multi-agent systems with continuous-time dynamics interacting through a very general class of monotonic continuous signal functions that covers estimation biases, approximation of discrete quantization, or state-dependent estimation. Our analysis reveals that, in the setup under consideration, synchronization equilibria are exactly the fixed points of the signal function. We also derive intuitive stability conditions based on whether the signal underestimates or overestimates the state of the agents around these fixed points. Moreover, we show that network topology plays a crucial role in asymptotic synchronization. These results provide interesting insights into the interplay between communication nonlinearity and network connectivity, paving the way for advanced coordination strategies in complex systems. I INTRODUCTION Information exchange is a fundamental aspect of both multi-agent systems (MAS) and networked control systems (NCS), as the agents’ dynamics are shaped by the information received from their neighbors. In many existing works on MAS, the interaction is considered to be perfect, i.e., the agents have access to the exact state of their neighbors [ 1 ] . This assumption simplifies the analysis but does not hold in real-world scenarios where communication is constrained due to bandwidth limitations, quantization effects, or noise. To address this limitation, several works [ 2 , 3 , 4 , 5 , 6 , 7 ] have investigated quantized communication models in which agents exchange only discrete or coarse information about their states. On the other hand, nonlinear consensus dynamics has also gained significant attention, particularly in models where agents interact through smooth and odd signals [ 8 , 9 , 10 ] . These approaches aim to capture more realistic scenarios where interactions are governed by nonlinear feedback mechanisms, often inspired by biological [ 8 ] or social [ 10 ] systems. However, these works primarily analyze behaviors around the neutral synchronization equilibrium using bifurcation theory. While this approach provides valuable theoretical insights, it does not fully capture the global behavior of the system, particularly in scenarios where multiple stable synchronization equilibria or complex nonlinear interactions arise. This limitation highlights the need for an analysis that goes beyond local bifurcations around the origin. In this paper, we present a framework for nonlinear consensus dynamics in which the interaction between agents is modeled by a general class of signals. By considering arbitrary monotonic, potentially non-smooth, Lipschitz-continuous signal functions, our analysis extends existing studies while preserving analytical tractability. This general setting encompasses classical linear consensus as well as continuous and piecewise approximations of discontinuous quantization, providing new insights on how communication nonlinearity influences the stability and convergence properties of the synchronization manifold in multi-agent systems. The key contributions of this paper are as follows: 1. We analyze a nonlinear consensus model in which agent interactions are described by monotonic Lipschitz-continuous signal functions. This formulation not only includes the traditional linear consensus model as a particular case but also accommodates a wide range of nonlinear continuous functions, as well as piecewise approximations of discontinuous quantization functions. 2. We characterize the equilibria of the system and show that synchronization equilibria correspond to the fixed points of the signal function. For that, we introduce intuitive notions of over- and under- estimation functions to analyze whether the synchronization occurs. Based on these concepts, we establish conditions for the stability of synchronized equilibria. We also highlight how the shape of the signal function influences the convergence behavior. Moreover, we provide conditions under which solutions converge to a locally stable synchronization equilibrium and analyze the corresponding basins of attraction. 3. We show that the network topology impacts synchronization by demonstrating that agents with similar neighborhood structures asymptotically synchronize independently of the signal function. The rest of the paper is organized as follows. In Section II , we formally define the problem and introduce the proposed nonlinear consensus model. Section III provides a detailed equilibrium and stability analysis. In subsection III-A and III-B , we analyze the invariance of the state space and the synchronization manifold, respectively. In subsection III-C , we analyze the synchronization equilibria of the dynamics induced by the communication signal, and in subsection III-D , we analyze the synchronization properties induced by the graph structure. We conclude the paper with some remarks and future work in Section IV . Notation In the following, we will denote by \mathbb{R} and \mathbb{R}_{\geq 0} the set of real numbers and the set of non-negative real numbers, respectively. For a vector \bm{x}\in\mathbb{R}^{N} , we denote by x_{i} the i -th component of \bm{x} . For a matrix \bm{A}\in\mathbb{R}^{N\times N} , we denote by a_{ij} the element of \bm{A} at the i -th row and j -th column. We denote by \bm{e}_{i} the i -th vector of the canonical basis of \mathbb{R}^{N} . We denote by \mathbf{1} the vector of \mathbb{R}^{N} with all components equal to 1 . Moreover, we denote by \mathrm{diag}(\bm{x})\in\mathbb{R}^{N\times N} the diagonal matrix with diagonal elements given by the vector \bm{x}\in\mathbb{R}^{N} . For a function f:\mathcal{X}\to\mathcal{X} , we denote \mathrm{Fix}(f)=\left\{x\in\mathcal{X}\mid x=f(x)\right\} the set of fixed points of f in \mathcal{X} . II PROBLEM FORMULATION We consider the classical multi-agent framework in which N individuals/agents belonging to the set \mathcal{V}=\left\{1,\dots,N\right\} , interact according to an undirected fixed graph \mathcal{G}=\left(\mathcal{V},\mathcal{E}\right) . We denote by \bm{A}\in\mathbb{R}^{N\times N} the symmetric adjacency matrix associated with the graph, i.e., a_{ij}=1 if (i,j)\in\mathcal{E} and a_{ij}=0 otherwise. We denote, by \bm{D}\in\mathbb{R}^{N\times N} the degree matrix , i.e., \bm{D} is the diagonal matrix with diagonal elements defined by d_{i}=\sum_{j=1}^{N}a_{ij} for all i\in\mathcal{V} . We also introduce \bm{L}:=\bm{D}-\bm{A}\in\mathbb{R}^{N\times N} the associated Laplacian matrix . The neighborhood of the agent i is denoted by \mathcal{N}_{i} and represents the set of agents that influence i according to the graph \mathcal{G} (i.e., j\in\mathcal{N}_{i}\Leftrightarrow(j,i)\in\mathcal{E} ). By definition, the cardinality of \mathcal{N}_{i} is d_{i} . Let us also recall that a path in \mathcal{G} is a finite sequence of edges (i_{1},i_{2}),(i_{2},i_{3}),\dots,(i_{p},i_{p+1}) such that (i_{k},i_{k+1})\in\mathcal{E} for all k\in\{1,\dots,p\} . Two vertices i , j\in\mathcal{V} are connected in \mathcal{G} if there exists a path in \mathcal{G} joining i and j (i.e., i_{1}=i and j_{p}=j ). The graph \mathcal{G} is connected if any two vertices are connected. The state of agent i is denoted by x_{i}\in\left[-1,1\right] and evolve according to the following dynamics \dot{x}_{i}=\frac{1}{d_{i}}\sum_{j=1}^{N}a_{ij}s(x_{j})-x_{i}, (1) where s:\left[-1,1\right]\to\left[-1,1\right] is a common non-decreasing Lipschitz-continuous function . The nonlinear dynamics ( 1 ) captures the imperfect communication between agents via a broad class of signal functions. For example, common choices include the linear map s(x)=x or the affine function s(x)=ax+b , which respectively model unbiased or biased estimations. One can also consider sigmoidal functions, such as s(x)=\tanh(x) , that are inspired by biological phenomena [ 8 ] or neuroscience [ 10 ] . Furthermore, ( 1 ) can describe the behavior of quantized communication, where discontinuities are inherent. Specifically, discontinuous quantization schemes, as in [ 2 , 5 , 4 ] , can be approximated using a continuous piecewise affine signal function wherein different linear segments connect at discontinuities. The qualitative behavior of the agents’ states remains unchanged since, as we will show later, synchronization occurs only at specific fixed points of the signal function. In practice, such