Emergence of solitary and chimera states in adaptive pendulum networks under diverse learning rules
R. Anand, V. K. Chandrasekar, R. Suresh · 2026
Read Emergence of solitary and chimera states in adaptive pendulum networks under diverse learning rules by R. Anand, V. K. Chandrasekar, R. Suresh (2026) 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-2603-10668.txt.
Introduction
R Anand, V K Chandrasekar, and R Suresh study how phase lag and adaptive learning rules shape states in a network of pendulum oscillators. Coupling strengths evolve under two biologically inspired rules: Hebbian adaptation and spike-timing-dependent plasticity. The Hebbian case yields cluster, solitary, and chimera states, while STDP gives splay and splay-cluster forms. The solitary state arises with no delays, nonlocal coupling, or outside kicks, driven purely by the phase-lag value. Phase diagrams from two incoherence measures map the transitions, with multistable regions confirmed by analysis.
Catalog record
| CREATOR | R. Anand, V. K. Chandrasekar, R. Suresh |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40719 |
| LICENSE | http://creativecommons.org/publicdomain/zero/1.0/ |
| RIGHTS | CC0 |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.10668
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Emergence of solitary and chimera states in adaptive pendulum networks under diverse learning rules R. Anand, V. K. Chandrasekar, R. Suresh Abstract We investigate the interplay between phase lag and adaptive learning rules in a network of identical pendulum oscillators, where the coupling strengths evolve dynamically in response to the oscillators’ states. Specifically, we examine two biologically inspired adaptation mechanisms, Hebbian and spike-timing-dependent plasticity (STDP), and their influence on the emergence of collective dynamical patterns. Under Hebbian adaptation, the network exhibits a wide range of organized behaviors, including two-cluster, solitary, multi-antipodal, and chimera states. In contrast, STDP coupling induces splay, splay-cluster, and splay-chimera configurations. Importantly, we find that the solitary state arises spontaneously in this adaptive network without requiring delays, nonlocal coupling, or external perturbations; instead, it is induced purely by variations in the phase-lag parameter. To the best of our knowledge, such delay-free and symmetry-preserving emergence of solitary behavior has not been reported previously in adaptive oscillator systems. To systematically characterize the resulting dynamical transitions, we employ two complementary incoherence measures based on the local standard deviation of (i) time-averaged frequencies and (ii) instantaneous phases across spatial bins, enabling the construction of detailed two-parameter phase diagrams. Analytical stability analysis of the two-cluster state shows strong agreement with numerical simulations, revealing regions of pronounced multistability. These findings establish adaptive pendulum networks as a minimal yet powerful framework for studying self-organized synchronization, chimera formation, and multistable transitions driven by diverse adaptation mechanisms. 1 Introduction The study of complex networks provides a fundamental framework for understanding collective dynamics observed across a variety of natural and engineered systems, from neuronal assemblies to power grids and social structures [ 1 , 2 ] . Research on network dynamics has traditionally focused on the emergence of synchronization, wherein individual oscillators evolve coherently [ 3 ] . However, subsequent studies, particularly in neural networks subject to stochastic noise [ 4 ] , have revealed that large-scale oscillatory behavior can spontaneously emerge even when individual elements remain desynchronized. Such emergent collective dynamics have been explored under diverse conditions, including time delays, heterogeneity among oscillators, and time-dependent coupling strengths [ 5 , 6 , 7 ] . Within this context, adaptive networks, systems in which the connection topology evolves concurrently with the nodal dynamics, have attracted substantial attention [ 8 , 9 , 10 , 11 , 12 , 13 ] . Adaptation mechanisms enable feedback between network structure and dynamics, leading to self-organized synchronization and other rich dynamic phenomena. For example, slow adaptation relative to the intrinsic timescale of the oscillators allows the use of singular perturbation analysis, revealing how long-term plasticity shapes collective states [ 13 ] . Adaptive dynamical networks are known to exhibit a variety of self-organized patterns, including multi-cluster synchronization, chimera states, and hierarchical frequency groupings, arising from the interplay between state evolution and topological adaptation [ 14 , 15 , 16 , 17 , 11 ] . Gradual adaptation is particularly crucial for stabilizing multi-cluster configurations, where transient formations such as double-antipodal clusters strongly influence long-term dynamics [ 18 ] . These networks can sustain partially synchronized or hierarchically organized states, such as splay, antipodal, and double-antipodal patterns, through a delicate balance between coupling evolution and network structure [ 19 ] . Similar mechanisms have been observed in adaptive systems ranging from Morris–Lecar neuronal models [ 20 ] to phase oscillators with dynamic couplings [ 21 , 22 , 23 , 18 ] . Recent investigations have further revealed heterogeneous nucleation phenomena [ 24 , 25 ] and frequency-dependent synchronization shifts in multilayer adaptive networks [ 26 ] . Such systems often feature slowly evolving coupling variables, which give rise to unique dynamical phenomena, including Canard cascades [ 27 ] . The formulation of adaptation rules is largely inspired by neuronal learning principles, particularly those describing the pre- and postsynaptic timing of neuronal firing [ 28 , 29 , 30 , 31 , 32 ] . Among these, Hebbian and spike-timing-dependent plasticity (STDP) rules remain the most extensively studied. Hebb’s postulate, summarized as “neurons that fire together wire together” [ 33 , 34 , 35 ] –implies that coupling strengths increase for in-phase firing and weaken for antiphase activity. These principles form the foundation for long-term memory formation and neuromorphic computing architectures [ 36 , 37 , 38 , 39 ] . Understanding how learning rules interact with phase delays is therefore essential for describing the emergent behaviours of adaptive systems [ 40 , 41 ] . Prior studies have reported a wide range of collective patterns, including stable phase clusters, antipodal clusters, chimera states, and explosive synchronization, in adaptive or multiplexed oscillator networks [ 19 , 42 , 43 , 44 ] . However, these works primarily focused on first-order phase oscillators or networks where solitary states appear only under nonlocal coupling or delay-induced effects [ 45 , 41 ] . In contrast, the present work investigates, for the first time, the collective dynamics of identical pendulum oscillators governed by adaptive coupling with an intrinsic phase lag, in the absence of any restoring force. We demonstrate that the system supports a wide spectrum of novel dynamical states: two-cluster (TC), solitary (SS), multi-antipodal (MAC), chimera (CHI), splay (SP), splay-cluster (SPC), splay-chimera (SPCHI), and incoherent (INC) states depending on the phase lag and adaptation rule. Importantly, we find that solitary states emerge in this globally coupled adaptive network without requiring delays, nonlocal coupling, or external forcing; instead, they arise purely from variations in the phase-lag parameter. To systematically characterize these dynamical behaviors, we employ two independent incoherence measures based on the local standard deviation of (i) time-averaged frequencies and (ii) instantaneous phases across spatial bins. These measures enable the construction of detailed two-parameter phase diagrams that clearly delineate the transitions between the observed dynamical regimes. We also provide an analytical stability analysis of the two-cluster state, showing strong agreement with numerical results and highlighting regions of multistability within the adaptive pendulum network. Figure 1: Evolution of coupling strengths under different adaptation rules characterized by the control parameter \beta . (a) Hebbian adaptation for \beta=-\pi/2 , (b) STDP for \beta=0 , and (c) anti-Hebbian adaptation for \beta=\pi/2 . The remainder of this paper is organized as follows: Section 2 presents the model formulation of the adaptive pendulum network. Section 3 discusses the dynamical states observed under different adaptation rules. In Sec. 4 , we analyze dynamical transitions and characterization measures using one- and two-parameter phase diagrams. Finally, Sec. 5 summarizes the main findings and their implications. 2 Model A key extension of the Kuramoto model to pendulum networks involves incorporating a second-order inertial term to capture the oscillators’ rotational dynamics [ 46 , 47 ] . This modification was originally developed within an adaptive framework to explain the precise synchronization of flashing in Pteroptyx malaccae fireflies, which achieve near-zero phase differences through slow adaptation [ 8 ] . Similar adaptive mechanisms have since been applied to model diverse biological, chemical, and technological systems, including power grid dynamics [ 48 , 49 , 50 ] . In the present study, we consider a globally coupled network of N identical pendulum oscillators, where each node interacts with every other node through adaptive coupling. The dynamics of