同步对于复杂网络化系统的稳定性与协调运行至关重要。牵制控制通过选择性地控制部分节点，为提升网络同步能力提供了一种可扩展的解决方案。然而，现有的策略存在关键局限性：基于启发式中心性的方法缺乏与同步动力学的直接联系，而基于谱分析的方法虽然有效但计算成本高昂。为解决这些挑战，我们提出了一种基于扰动优化的策略（PBO），该策略动态评估每个节点对拉普拉斯矩阵的谱影响，从而在显著降低计算成本（复杂度为O(kM)）的情况下实现更好的同步能力。大量实验表明，所提出的策略在同步能力、收敛速度以及应对节点失效的牵制鲁棒性方面优于传统策略。值得注意的是，在所有测试的实证网络及一些生成网络中，PBO明显优于暴力贪婪策略，证明其能够避免局部最优解并适应复杂的连接模式。本研究建立了网络同步能力与收敛速度之间的理论关系，为大规模复杂网络的有效同步策略提供了新见解。

Synchronization is essential for the stability and coordinated operation of complex networked systems. Pinning control, which selectively controls a subset of nodes, provides a scalable solution to enhance network synchronizability. However, existing strategies face key limitations: heuristic centrality-based methods lack a direct connection to synchronization dynamics, while spectral approaches, though effective, are computationally intensive. To address these challenges, we propose a perturbation-based optimized strategy (PBO) that dynamically evaluates each node's spectral impact on the Laplacian matrix, achieving improved synchronizability with significantly reduced computational costs (with complexity O(kM)). Extensive experiments demonstrate that the proposed method outperforms traditional strategies in synchronizability, convergence rate, and pinning robustness to node failures. Notably, in all the empirical networks tested and some generated networks, PBO significantly outperforms the brute-force greedy strategy, demonstrating its ability to avoid local optima and adapt to complex connectivity patterns. Our study establishes the theoretical relationship between network synchronizability and convergence rate, offering new insights into efficient synchronization strategies for large-scale complex networks.