Multi-relational network layer node immunization strategy based on multi-particle random walk

With the rapid development of science and technology, the ways in which rumors and diseases spread have also undergone significant changes, profoundly impacting people’s lives. In this context, finding effective immunization methods within complex networks has become a topic of considerable interest. Although research on single-layer networks has yielded mature results, there is still a need for further exploration of immunization strategies in multi-layer networks. This paper first extends the multi-particle random walk algorithm to multi-layer networks and proposes a multi-particle random walk immunization strategy for layer nodes (MLRWI). Given that in the real world, particle walks typically exhibit bias—favoring the selection of more important or closer target nodes—we further propose a biased multi-particle random walk immunization strategy for layer nodes (BMLRWI). Experimental results on four real-world datasets show that BMLRWI achieves good performance compared to other immunization strategies without requiring global information. Additionally, the experiments demonstrate that immunization strategies that consider bias significantly outperform the unbiased MLRWI in terms of effectiveness.