<p>Fair Influence Maximization (FIM) aims to ensure equitable dissemination of influence across diverse user groups while maximizing the overall spread within a social network. However, existing methods often ignore overlapping communities and rely on a single transformation strategy, limiting diffusion fairness and adaptability. To address these challenges, we develop FEOCO (Fair Evolutionary Overlapping Community Optimization), a multi-objective evolutionary framework that seamlessly integrates overlapping community detection with multi-transformation optimization. By employing a multi-transformation optimization framework, FEOCO simultaneously optimizes multiple transformation models as parallel tasks using a unified evolutionary solver. The potential relationships between transformations are estimated based on the degree of overlap among individuals from different populations. Meanwhile, FEOCO ensures fair consideration of both non-overlapping and overlapping nodes when identifying key seed nodes. Furthermore, FEOCO formulates the various network transformations as parallel optimization tasks that are iteratively refined through the evolutionary mechanisms of the Learner Performance-based Behavior (LPB) algorithm. Extensive experiments on real-world networks demonstrate that FEOCO achieves superior performance in both fairness and influence spread, outperforming state-of-the-art baselines. Specifically, FEOCO effectively leverages the potentially transferable knowledge across multiple transformations, resulting in an 8.4% improvement in average influence spread compared to current state-of-the-art techniques.</p>