Diffusion inspires selection of pinning nodes in pinning control

• Large-degree pinning nodes perform badly in real-world networks.
• Pinning node selection is transferred into multi-spreaders in information diffusion.
• Poor performance of large-degree selection is due to overlapping influence of pinning nodes.
• An effective pinning node selection is proposed to enhance pinning controllability.

The outstanding problem of controlling a complex network via pinning is related to network dynamics and has the potential to master large-scale real-world systems as well. This paper addresses the heart issue about how to choose pinning nodes for pinning control, where pinning control aims to control a network to an identical state by injecting feedback control signals to a small fraction of nodes. We explore networks’ controllability from not only mathematical analysis, but also the aspects of network topology and information diffusion. Then, the connection between pinning control and information diffusion is given, and pinning node selection is transferred into multi-spreader problem in information diffusion. Based on information diffusion, a heuristic method is proposed to select pinning nodes by optimizing the spreading ability of multiple spreaders. The proposed method greatly improves the controllability of large practical networks, and provides a new perspective to investigate pinning node selection.