考虑小团体行为特征的客轮疏散建模及管控策略优化研究

Evacuation is an important mean to deal with emergencies, of which organization and management is of vital importance to crowd safety. There are many small groups such as relatives and friends on the passenger ship, their movement and other behaviors have a significant impact on the overall evacuation of the crowd. However, the evacuation decision-making mechanism that determines the temporal and spatial distribution of small groups is still unclear so far. In view of this, considering the limited moving capacity of passengers on ships of special status (e.g., pitching and rolling) and evacuation guidance of ship crew, we will carry out small group walking experiment and crowd evacuation behavior experiment to explore cooperative moving rules of small groups as well as decision-making rules of the individual and small group with the influence of evacuation guidance. Based on the above pedestrian behavior rules and individual moving characteristics on ships of special status, an Agent-based crowd evacuation model will be established. By changing the composition of crowds, facilities availability, evacuation guiding modes and other conditions, typical ship scenarios will be established, and the change rule of crowd flow efficiency and traffic flow of key facilities under different scenarios will be analyzed. Then a number of methods to optimize crowd flow can be proposed. Moreover, we will construct a bi-level optimization framework based on heuristic algorithm and pedestrian evacuation simulation model, by which the optimized zonal evacuation guidance strategies for varied passenger ship scenarios will be calculated and obtained. The research results of this project are of great significance to understand the dynamics evolution law of passenger ship evacuation, which can provide scientific basis for the emergency management of cruise ship evacuation.

疏散是应对突发事件的重要手段，合理组织与高效管控对人员安全影响重大。客轮上亲友等小团体众多，结伴移动等行为会显著降低人群整体疏散效率，但目前决定小团体时空分布的疏散决策机制仍不明确。鉴于此，考虑船体倾斜等特殊状态对人员的行走限制以及船员的引导作用，本项目开展小团体人员行走实验和人群疏散行为实验，分析并建立小团体协同运动规则以及疏散诱导影响下独立个体、小团体的决策规则，结合船体特殊状态下人员的运动特征，构建基于主体的客轮人群疏散模型，考虑不同人群构成、疏散设施可用性、疏散诱导等条件设定若干典型客轮疏散场景，进行动力学分析并总结不同场景下人群疏散效率与关键设施通行流量的变化规律，提出优化人群流动的方法，建立基于启发式算法与行人疏散仿真模型的双层优化框架，获取针对典型客轮场景的分区疏散引导优化策略。项目成果对于深入理解客轮疏散动力学演化规律有重要意义，可以为客轮应急管理工作提供科学依据。

作为移动的水上建筑，客轮人员密集、结构复杂、布局紧凑，一旦发生突发事件，可能造成严重的人员伤亡与财产损失，而高效组织人员疏散是应对突发事件的重要手段。客轮疏散既具备一般建筑疏散的复杂性，还具有显著特点：船舶遇到触礁等突发事故可能呈现倾斜状态，对人员的运动行为有较大影响；小团体乘客较多，其运动及决策行为与独立个体有差异；船员熟悉客轮环境，可引导乘客疏散。.项目组针对客轮疏散特点，总结分析了船体不同倾斜、摇晃状态下人群运动规律，设计并实施了行人运动实验，提取人员运动轨迹，获取了不同类型障碍物对人员运动行为的影响规律，建立了基于启发式规则的行人运动模型，同时利用神经网络模型，由实验数据驱动实现了行人轨迹的精确预测。然后在复杂建筑布局内设计了可控的小团体疏散实验，并基于虚拟现实平台开展了个体疏散决策实验，分析了个体及小团体的出口、路径选择等疏散决策模式，运用博弈论建立了疏散行为模型。在总结个体、小团体运动及决策行为规律的基础上，建立基于主体的客轮人群疏散模型，将疏散模拟结果与演习数据对比进行了模型验证。开展客轮疏散动力学分析，获取小团体行为对客轮疏散效率以及关键区域通行流量的影响规律。考虑昼夜差异、疏散人员总数量、船舶倾斜角度、楼梯等疏散设施的可用性等因素，构建典型仿真场景，并以各场景的模拟结果作为驱动，开发客轮人员疏散时间快速预测方法。在此基础上，建立了基于遗传算法与行人疏散仿真模型的双层优化框架，获取针对复杂空间布局的分区疏散优化策略。研究成果能为客轮等复杂空间人群安全管控提供理论参考和决策支持。.在项目执行期间，项目组与本领域国内外学者开展交流与合作，在国际国内重要学术期刊总计发表了高水平学术论文11篇，其中SCI检索5篇，获批软件著作权2项。培养硕士研究生10名，其中7名已获硕士学位，建立并形成了稳定的研究团队，围绕行人交通和人员疏散动力学开展深入的科学研究。