Abstract: By investigating the invulnerability of semi-directed wireless sensor networks, this paper provides a theoretical basis for formulating effective invulnerability strategies. The paper uses the grey system theory to characterize the widespread incomplete information in the network. By applying the generating function method, it successfully derives the critical removal ratio and the size of the giant connected component of semi-directed random networks with arbitrary degree distribution. The topologies of semi-directed wireless sensor networks are described by simulation, and the invulnerability analysis of semi-directed wireless sensor networks is carried out, with the Barabási-Albert (BA) scale-free network taken as a typical case for simulation analysis. The results indicate that the intentional dynamic attack has obvious advantages over the random attack and the intentional static attack in attack efficiency. This conclusion provides an important reference for formulating strategies to improve the invulnerability of wireless sensor networks.

Key words: wireless sensor networks, semi-directed networks, generating function, invulnerability