A network is a topology consisting of nodes and edges connecting nodes, where nodes interact with each other through edges for information and to fulfil the functions of the network. When the network topology is attacked by different types of attacks, the network’s information perception ability, execution ability and coordination ability will be affected. Dynamic reconfiguration of the network based on node visual perception is an effective method to ensure reliable communication in the network. Therefore, discovering influential node neighbours to connect through the visual attention mechanism is beneficial to enhance the network’s destruction resistance and robustness. In this work, a swarm neighbour selection model based on visual attention mechanisms and a network connectivity model considering geographic distance and saturation values are used to perform dynamic reconfiguration of the two swarm networks to solve the problem of isolated nodes accessing the network. Meanwhile, a swarm network robustness assessment method considering dynamic reconfiguration is proposed to evaluate the robustness of networks with different reconfiguration methods and compare and analyse them. The results show that the network reconfiguration method based on the visual attention mechanism is effective and feasible for these two swarm networks, and has outstanding advantages and performance in terms of robustness and destruction resistance compared with other reconfiguration methods.
Wu et al. (Wed,) studied this question.