基于群集运动大数据的网络群体智能理论与算法研究

When organisms that follow simple rules appear in the form of coupled networked groups, they often exhibit amazing complexity and coordination. This is called “networked swarm intelligence”, and it has become an important research direction of the new generation of artificial intelligence. Understanding the collective movement mechanism of natural organisms and active matters is the key way to study the theory and application of networked swarm intelligence. For instance, revealing the behavioral rules of immune cells is significant for the development of immunology and the development of new cancer drugs. This project takes the data of collective behavior of immune cells obtained by living imaging technology as the starting point, conducts research on the analysis strategy of spatiotemporal characteristics of collective motion, swarm intelligence algorithm based on machine learning, and the data-driven modeling method for networked swarm behavior. The project aims to propose systematic approaches to analyze the self-organizing rules of swarm intelligence, explore the causal relationship between individuals, construct the interaction network and the coupled system dynamic model, and reveal the mechanism of swarm intelligence of this kind of immune cells. Finally, it will propose the networked swarm intelligence model combined with the spatio-temporal characteristics of the system, and build the software simulation platform for collective behavior. The results of the research project are potential to provide the support of data mining and modelling analysis for applications such as robot cluster, UAV formation, panic crowd evacuation, and system biology.

遵循简单规则的生物以耦合的网络化群体形式出现时，常表现出惊人的复杂性与协同性，这种现象被称为“网络群体智能”，已成为新一代人工智能的重要研究方向。理解自然界生物与活性物质的群集运动机理是网络群体智能的理论与应用研究的关键途径，例如，揭示免疫细胞群体行为规律对免疫学发展以及新型癌症药物研制具有重要意义。因此，本项目以基于活体成像技术获取的免疫细胞群体行为数据为切入点，开展群集运动大数据时空特性分析策略研究，基于机器学习的群体智能算法研究，以及数据驱动的网络群体行为建模方法研究。项目旨在提出分析群体智能自组织规律的系统化方法，挖掘个体间因果关联，构建互动网络与耦合系统动力学模型，揭示此类免疫细胞的群体智能规律。最终提出融合系统时空特性的网络群体智能模型，并搭建群体行为软件仿真平台。项目成果预计将为机器人集群、无人机编队、恐慌人群疏散、系统生物学等领域的应用提供数据挖掘与建模分析的支持。

遵循简单规则的生物以耦合的网络化群体形式出现时，常表现出惊人的复杂性与协同性，这种现象被称为“网络群体智能”，已成为新一代人工智能的重要研究方向。理解自然界生物与活性物质的群集运动机理是网络群体智能的理论与应用研究的关键途径，例如，揭示免疫细胞群体行为规律对免疫学发展以及新型癌症药物研制具有重要意义。因此，本项目以基于群体行为数据为切入点，开展群集运动大数据时空特性分析策略研究，基于机器学习的群体智能算法研究，以及数据驱动的网络群体行为建模方法研究。项目旨在提出分析群体智能自组织规律的系统化方法，挖掘个体间因果关联，构建互动网络与耦合系统动力学模型，揭示一般的群体智能规律。最终提出融合系统时空特性的网络群体智能模型，并搭建群体行为软件仿真平台。项目成果预计将为机器人集群、无人机编队、恐慌人群疏散、系统生物学等领域的应用提供数据挖掘与建模分析的支持。