The purpose of influence maximization in a cascade model is to extricate a group of hubs from underlying network that can maximize propagation. Greedy methods may be used to perform influence maximization. However, as it has ahigh computational cost, it becomes inefficient. Centrality-based heuristic methods can be considered as a replacement. But it gets trapped in local optima. To address these issues, a framework has been introduced which uses community detection, tries to find the most significant nodes from each community using closeness centrality. Then, it employs Coral Reefs Optimization to determine the best solutions, i.e., the most important nodes that maximize propagation. Adaptive β-Hill Climbing has been used. It acts as a local search to avoid getting caught in local optima. A few datasets were used to evaluate the proposed framework’ scalabilities. Moreover, the outcomes were compared to several state-of-the-art techniques based on 2-hop influence spread and Monte-Carlo simulations.