Description
This project uses agent-based models to simulate how people with different interests gather into small groups based on shared interests, with the aim of optimizing modularity maximization for network community detection using agent-based models in the NetLogo platform. Since the computational power required by traditional methods in dealing with large networks grows nonlinearly with the number of elements in the matrix, and if a node or link is added or arbitrarily deleted, it will make the network structure change and require recalculation. Therefore, the complexity and computational effort of traditional methods for community detection in dynamic networks is extremely high. To solve this problem, we use a multi-agent system in which agents represent nodes in a network that form a network with community distribution through their interaction and adaptive behavior. In this study, we simulate real-world social behaviors by giving each agent a different number of attributes to observe the changes in the network structure through an agent-based model. Our model can efficiently perform community detection in dynamic network environments and, at the same time, instantly reflect the impact of the addition or removal of new agents on the network structure. By comparing with the traditional method in R, we verify the feasibility of the model in terms of community detection accuracy under different network sizes and complexity. The deviation from the agent-based model for community detection does not vary linearly with the number of nodes in R, as compared with the traditional method, and the average error is around 5 percent. Moreover, the number of interactions required to stabilize the network community structure in the agent-based model grows linearly with the number of attributes. This study provides a new technical tool for dynamic social networks and a new perspective for understanding and predicting social network behavior.| Period | 1 Jan 2024 → 24 Dec 2024 |
|---|---|
| Degree of Recognition | International |