Fractional Atangana–Baleanu Dynamics on Signed Graphs: Consensus, Synchronization, and Community Detection
This paper develops a fractional-order dynamical framework on signed graphs by combining the Atangana–Baleanu (AB) fractional derivative in the Caputo sense with the signed Laplacian operator. The AB operator’s non-local, non-singular Mittag-Leffler kernel effectively captures memory-dependent processes and non-Markovian behavior in complex networks. We study the system ${}^{ABC}D_t^\alpha \mathbf{x}(t) = -L_\Sigma\mathbf{x}(t) + \mathbf{f}(\mathbf{x}(t))$ and establish rigorous theoretical results including existence and uniqueness of solutions, asymptotic stability via Mittag-Leffler decay, and bipartite consensus on structurally balanced graphs. We also investigate synchronization properties and noise robustness of the fractional system. Signed graph coloring is used as a motivating framework to interpret the discrete state assignments that emerge from the continuous fractional dynamics at steady state: nodes connected by positive edges reach agreement (same color class), while those connected by negative edges diverge (different color classes). Numerical experiments on social, biological, and synthetic networks demonstrate that fractional dynamics with order $\alpha \approx 0.7$ achieve faster consensus, more accurate community detection, and superior noise robustness compared to integer-order models. These results open new application opportunities in social networks, biological systems, and multi-agent coordination.
Authors
- Pie Desire Ebode Atangana (ORCID: https://orcid.org/0000-0002-4251-8719)
Institutions
- Université de Yaoundé I (CM)
Publication Details
- Journal
- African Journal for Industrial and Applied Mathematics
- Published
- 2026-11-18
- DOI
- https://doi.org/10.4208/ajiam.202x-xxxx
- Primary Topic
- Fractional Differential Equations Solutions
- Type
- article
- Field-Weighted Citation Impact
- 0.00