2103.09625
Asymptotic and finite-time cluster synchronization of coupled time-varying delayed neural networks
Juan Cao, Fengli Ren, Dacheng Zhou
correctmedium confidence
- Category
- math.DS
- Journal tier
- Specialist/Solid
- Processed
- Sep 28, 2025, 12:56 AM
- arXiv Links
- Abstract ↗PDF ↗
Audit review
The paper proves two results: (i) asymptotic cluster synchronization under pinning impulsive control via a Lyapunov–Razumikhin/Halanay inequality and an impulse contraction bound, with rate λ/2, and (ii) finite-time cluster synchronization under a hybrid controller, via a Lyapunov functional whose derivative satisfies V̇ ≤ −c V^α (0<α<1). The candidate solution reproduces the same Lyapunov choices, the same Young/Schur-complement manipulations, the same Razumikhin condition with q ≥ γe^{λτ}, the same impulse inequality V(t_k^+) ≤ η_k V(t_k), and the same finite-time argument (Bhat–Bernstein), matching the paper’s steps almost line-by-line. See Theorem 1 (conditions (8)–(11) and proof steps leading to (18)–(19) and the jump bound (20)–(22), concluding rate λ/2) and Theorem 2 (controller (23), derivative (26), and the application of the finite-time lemmas) in the PDF . Minor issues: in Theorem 1, (11) appears to contain a typographical slip (μ+ευ−(σ+λ)<0) that should read μ+νq−(σ+λ)<0, consistent with the proof using c=μ+νq; and both the paper and model rely on a noncausal term ∫_{t}^{t1} in the hybrid controller, which raises implementability concerns although it does not invalidate the formal Lyapunov proof.
Referee report (LaTeX)
\textbf{Recommendation:} major revisions
\textbf{Journal Tier:} specialist/solid
\textbf{Justification:}
The manuscript adapts well-known Lyapunov–Razumikhin/Halanay techniques and finite-time analysis to a cluster-synchronization setting with time-varying delays and pinning impulses. The proofs are largely sound and aligned with standard results, but there is a typographical error in a key inequality and several steps (bounds for non-symmetric coupling, explicit extraction of constants from the LMI) would benefit from clarification. Most critically, the hybrid controller contains a noncausal integral of future error, which is not implementable as stated. Addressing these would substantially strengthen correctness, clarity, and practical value.