Quantum Particle Swarm Optimization Algorithm Based on Social Learning for Solving the Influence Maximization Problem

Quantum Particle Swarm Optimization (QPSO) is an heuristic algorithm used for optimization. When solving the influence maximization problem, the algorithm relies solely on the global best position and the average best position of the population for updates. This approach lacks diversity guidance, making it prone to falling into local optima, which leads to a “two steps forward, one step back” phenomenon—most dimensions advance, while a few dimensions degrade. In this paper, a Quantum Particle Swarm Optimization Algorithm based on Social Learning (SLQPSO) is proposed. By establishing a dimension-based knowledge base and storing different learning objects, a fitness-distancebased environmental metric operator is developed. This operator effectively reflects the environment of the particle swarm, balancing exploration and exploitation, and helps solve the “two-steps-forward, one-step-back” phenomenon. Experiments on six real-world social networks under the Independent Cascade (IC) model demonstrate that the proposed algorithm can effectively identify important node sets.