Discrete Whale Optimization Algorithm for Influence Maximization in Social Networks

ABSTRACT Influence Maximization (IM) is a core problem in social network analysis. It aims to identify a set of a fixed number of key nodes, known as the seed set, to maximize the eventual spread of information or influence under a given diffusion model. Research on influence maximization contributes to a deeper understanding of social networks and the mechanisms underlying viral marketing. Owing to their high computational cost and substantial execution time, traditional influence maximization methods often exhibit limited effectiveness when applied to real‐world social networks. Moreover, even heuristic approaches to influence maximization tend to achieve only modest performance gains despite their reduction in time complexity. This makes it an ongoing challenge to develop low‐cost algorithms that are both efficient and adaptable to diverse social networks. To address this issue, we propose a discretized whale optimization algorithm (DWOA) and integrate community detection to partition the network, thereby optimizing seed node selection and improving influence spread. The proposed method leverages the search capability of WOA and introduces a One‐Hop Replacement Strategy to reduce search blindness and strengthen population exploration. Meanwhile, a candidate‐node‐based random population initialization scheme is employed to accelerate convergence and lower the overall computational overhead. We evaluate the proposed DWOA on eight real‐world social networks using the Independent Cascade (IC) model. The experimental results show that DWOA matches or surpasses competing metaheuristic methods across five performance metrics, while delivering substantially superior results compared to heuristic baselines.