Multi-hop centralized seeding strategy for influence maximization in information limited networks

Maximizing influence in complex social networks represents a fundamental challenge in the field of information dissemination, particularly in real-world scenarios where network topology and propagation mechanisms are often unknown. Traditional methods, which rely on global network information, face significant limitations such as high data acquisition costs and substantial computational complexity, rendering them inadequate for dynamic and unknown networks. To address these challenges, we propose a Multi-Hop Strategy (MHS) framework based on multi-hop perception, which overcomes the limitations of one-hop approaches by integrating the friendship paradox in social networks with human perceptual capabilities. The developed method dynamically selects high-influence seed nodes using local perceptual data, eliminating the need for global topological information. To validate the effectiveness of the proposed framework, extensive experiments are conducted on 12 real world networks spanning diverse domains. Our results show that MHS significantly outperforms four random baseline strategies and the one-hop strategy across most networks, while exhibiting strong robustness to network sparsity and heterogeneity.