Although a large number of studies have verified and explained the controllability of complex networks in real life and nature, there is a deficiency of accurate control strategies based on the proposed theory of network controllability. Here, we propose a new dimension reduction method, which firstly decouples the $N$-dimensional interdependent system into $N$ independent systems, then re-couples them into one state space. The tool can help predict the state of individual nodes, explore the behavior pattern of different dynamic models in the network, and quantify the responses of the network states in terms of its own structure and external disturbances. The results show that for nonlinear dynamical models with biochemical dynamics, birth–death processes, regulatory dynamics and epidemic processes on Scale-Free and Erd$\ddot{o}$s–R$\acute{e}$nyi networks, the activity of the target node or target node set can be accurately reached by controlling the behavior of some nodes with our framework.

尽管大量的研究已经验证和解释了复杂网络在现实生活和自然界中的可控性，但是基于所提出的网络可控性理论的精确控制策略存在不足。在这里，我们提出了一种新的降维方法，它首先将$ñ$维相互依存系统成$ñ$独立的系统，然后将它们重新耦合到一个状态空间中。该工具可以帮助预测单个节点的状态，探索网络中不同动态模型的行为模式，量化网络状态对自身结构和外部干扰的响应。结果表明，对于具有生化动力学、出生-死亡过程、调节动力学和流行过程的非线性动力学模型，Scale-Free 和 Erd$\ddot{○}$s–R$\acute{e}$nyi 网络，通过我们的框架控制一些节点的行为，可以准确地达到目标节点或目标节点集的活动。