Abstract
The essence of the influence maximization (IM) problem is how to identify the set of seed nodes so that the node numbers ultimately affected in the network reach the maximum under a certain spreading model. In the field of influence maximization research, the investigation of seed nodes identifying algorithms is a hot yet challenging work. Although conventional greedy algorithms and heuristic algorithms have high performance, their efficiency remains a challenge when applied to large-scale social networks. In recent years, swarm intelligence-based optimization algorithms have seen increasing application in addressing this problem, with notable improvements in performance. However, the efficiency of these swarm intelligence-based algorithms still needs to be improved in large-scale social networks. Based on this issue, a parallel discrete crow search algorithm (PDCSA) designed for parallel computing is proposed. Based on the evolution characteristics, PDCSA makes full use of the efficiency advantage of parallel computing to improve the time efficiency of solving IM problems.The results of experiments conducted on six datasets show that PDCSA achieves performance comparable to state-of-the-art algorithms, with the added advantages of high efficiency and robustness.