<p>The influence maximization aims to identify a subset of highly influential nodes as seeds to achieve the most widespread dissemination of targeted information within a social network or more. Individuals in practice inherently participate in multiple social networks, demonstrating flexible interaction behaviors and intricate influence propagation dynamics. However, existing research has predominantly concentrated on simplified single-layer networks, and the seed identification approaches often overlook the understanding of inter-layer propagation relationships. This paper proposes a novel layer weight-driven evolutionary optimization (LWEO) to address the influence maximization problem in a more effective way in multilayer networks. The layer weights of the network are quantified and combined with a fitness evaluation function, which is designed specifically to depict the characteristics of multilayer networks, to assess candidate seed sets during the evolution process. To drive the candidate sets toward the optimal solution, node centrality indices are incorporated into the evolutionary optimization. Meanwhile, a local search strategy is conceived to replace the suboptimal influential nodes based on their centrality and local influence rankings. Extensive experiments conducted on both synthetic and real-world networks validate the feasibility and effectiveness of the proposed algorithm. The results demonstrate that the LWEO achieves an average improvement of 6.6% in influence propagation in small networks and 14.3% in medium networks compared to state-of-the-art algorithms.</p>