MHPR: Identification of Key Nodes in Multilayer Heterogeneous Combat Network Based on Improved PageRank Algorithm

Network defense is crucial in ensuring the stable operation of the infrastructure network and maintaining the combat capability of the combat network. It also plays a vital role in improving the network's survivability against malicious attacks and random failures. This paper proposes a key node identification method based on an improved PageRank algorithm to address the combat network defense problem, which is transformed into a defense problem of key nodes in multilayer heterogeneous networks. The attribute PageRank (APR) method is introduced, which comprehensively considers the influence of node attributes and topology information on key node identification. The combat network is modeled as a multilayer heterogeneous network, and the proposed Multilayer Heterogeneous network PageRank (MHPR) algorithm is designed to identify and defend key nodes, with the objective of maximizing the network defense capability while defending a small number of nodes. The performance of MHPR is evaluated on networks of varying scales, and the results show that it outperforms the traditional PageRank algorithm, with a 7.3%-8.6% improvement in cumulative normalized defense capability (ANDC). Furthermore, compared with other existing methods, MHPR achieves a 6.1 %-7.4% improvement in ANDC.