近年来大型活动举办越来越频繁。大型活动涉及人员众多，涉及部门繁杂，手机、微信、APP、安检设备、移动通讯车和应急平台等大量投入使用，针对其突发事件的安全管理至关重要。目前的安全管理以被动的人防为主，很难满足新形势下的安全管理需求。本项目面向大数据背景下的全景式管理机遇，聚焦于大型活动的公共安全管理与决策方法研究。基于多源异构数据融合分析方法，建立基于多维特征的重点群体识别与监控方法。综合考虑突发事件情景建模和多主体多目标的协同模式，构建基于博弈理论和组织行为学的Multi-Agent模型，提出数据驱动的近实时的情景推演方法。研究“数据-模型-知识经验”精准决策、“政府-组织-公众”多主体参与的协同决策和跨部门跨领域协同应对方法，建立深度融合微观、中观和宏观层次的全景式安全管理与决策方法。项目预期能够对常态和非常态下大型活动的安全管理提供强有力的支持和帮助，丰富我国公共安全管理决策理论体系。

Recently, large-scale events are organized more and more frequently in China. Cell phone, WeChat, mobile communications vehicle and emergency platform system are widely used. It is very important for the safety management under emergency. Depending on the passive prevention, current safety management is difficult to satisfy the requirements under new situation. Facing opportunities of the panoramic management in Big Data, this project focuses on the methods for safety management and decision-making of large-scale events. Based on fusion and mining of multi-source and heterogeneous data, we will study the methods for key group identification and control. Taking the scenario building and collaborative mode of multi-agent and multi-object into account, we need to build the Multi-Agent model with game theory and organizational behavior, and propose the method for data-driven and up to real time scenario deduction. By combining the ‘data-model-knowledge’ precision decision, collaborative decision making from multi-agencies (including government, organization, public) and inter-department and inter-disciplinary, we put forward the holistic management and decision-making method that involve the micro, meso and macro levels. The project is expected to provide support for safety management of large-scale events under normal and emergency circumstances, and will enrich a series of theories on public safety management.