Multi-Objective Framework Boosts Fairness and Information Spread Across Overlapping Social Communities

A newly developed computational framework improves both equity and total influence spread across overlapping community structures without sacrificing overall diffusion performance, according to an investigation published in the Journal of Big Data by researchers Ye Wang, Yurong Zhang, and Liming Wang. The peer-reviewed study demonstrates that formulating influence diffusion from a multi-objective perspective allows network algorithms to achieve broad information distribution while actively mitigating unequal information access across distinct social groups.

Within social network analysis, influence maximization focuses on identifying an effective set of seed nodes capable of maximizing the spread of information throughout a complete network. Traditional influence maximization approaches have historically prioritized overall diffusion volume above all else. Under this conventional paradigm, algorithms concentrate almost entirely on expanding total reach, frequently neglecting whether the information is distributed equitably across the different communities that comprise the network.

This systematic neglect becomes particularly severe in social networks that feature overlapping community structures, where individuals routinely belong to more than one group at the same time. The authors observe that failing to account for these intersecting boundaries leads directly to biased diffusion dynamics. Over time, such biased dissemination produces unequal information access among different social groups, leaving certain populations systematically underserved by standard influence maximization methods.

To overcome these limitations, the research team introduced a comprehensive framework that addresses diffusion performance and fairness simultaneously rather than viewing them as competing goals. Specifically, the authors formulated the computational challenge as Multi-Objective Fair Influence Maximization in Overlapping Communities, designated as MOFIM-OC. Under this formulation, the optimization process jointly targets both influence diffusion performance and equity fairness across the network’s overlapping community structures.

The proposed MOFIM-OC approach explicitly incorporates overlapping community characteristics to better capture complex real-world social interactions. By integrating these structural attributes directly into the modeling process, the method ensures balanced and fair consideration of both shared nodes that participate in multiple groups and non-shared nodes that belong to only one. Additionally, the researchers introduced a multi-transformation mechanism that enhances search capabilities by establishing multiple representation spaces for analyzing diffusion dynamics.

To solve the resulting optimization problem efficiently, the authors developed a unified multi-objective framework that integrates evolutionary search strategies with transformation-based learning mechanisms. This hybrid algorithmic design substantially improves exploration capability throughout the optimization process while maintaining a careful balance between solution convergence and population diversity. Consequently, the integrated method enhances the overall quality and trade-off balance of the generated Pareto optimal solutions.

The researchers evaluated the proposed approach through extensive experimental testing conducted on several real-world social networks. The experimental findings demonstrated that the unified framework consistently outperforms state-of-the-art baselines in terms of both influence spread and fairness metrics. These performance results confirmed that combining overlapping community structures with multi-objective optimization and multi-transformation mechanisms significantly enhances diffusion equity across groups without compromising overall network influence spread.

Subject of Research: Fair influence maximization in overlapping community social networks

Article Title: Enhancing equity and influence spread in overlapping communities via multi-objective optimization and multi-transformation framework

Article References: Wang, Y., Zhang, Y., & Wang, L. (2026). Enhancing equity and influence spread in overlapping communities via multi-objective optimization and multi-transformation framework. Journal of Big Data. https://doi.org/10.1186/s40537-026-01600-8

Image Credits: AI Generated

DOI: 10.1186/s40537-026-01600-8

Keywords: influence maximization, overlapping communities, multi-objective optimization, multi-transformation framework, MOFIM-OC, diffusion equity, social networks, evolutionary search, transformation-based learning, Pareto solutions, seed nodes, information spread, community structures, biased diffusion, shared nodes, diffusion dynamics, Journal of Big Data