建筑物内行人流疏散动态演化博弈研究

发布时间：2018-11-26 19:56

【摘要】：随着人民生活水平的不断提高,城市化进程的不断加快,现代城市各类公共场所人口和资源不断集中,各种风险和非常规突发事件的威胁日益凸现。非常规突发事件引起的行人疏散过程中的拥挤踩踏等群体灾难性事故后果往往是很严重的,造成的社会影响也非常恶劣。人们对城市安全的重视程度越来越高。为保证人们的人身安全,开展对公共场所高密度行人群体疏散模拟仿真、风险评估和应急管理等多方面科学探索研究具有重要的意义。目前,疏散研究吸引了不同专业背景的学者,成为最具研究价值的课题之一。针对行人疏散本文主要围绕元胞自动机模型和演化博弈理论模型进行研究,主要内容包括以下几个方面:首先,基于元胞自动机模型,构建行人疏散过程中竞争行为与合作行为的收益参数,利用收益参数值来确定行人的移动规则,然后在不同系统规模、出口宽度、行人密度条件下进行模拟仿真,统计疏散时间。通过比较分析发现,在系统规模和出口宽度一定的条件下,行人疏散时间随行人密度的增加呈线性增长趋势;在系统规模和行人密度一定的条件下,行人疏散时间随出口宽度的增大呈负指数减小趋势;紧急疏散条件下高密度行人采取合作行为有利于提高疏散效率。其次,为了促使紧急疏散条件下高密度行人选择合作行为,提出惩罚机制下囚徒困境的元胞自动机模型。采用弱囚徒困境量化行人行为的收益,利用惩罚力度作为对行人选择竞争行为的惩罚,讨论Von Neumann邻域和Moore邻域中,行人竞争行为在不同初始状态下与惩罚力度的关系,从而进一步对行人密度分别为0.6和0.8的情况进行仿真。通过分析发现,在疏散过程中,行人行为的变化速度与邻域的个数有关,邻域数量越多,行人行为越不容易改变;惩罚力度越大,选择竞争行为的行人承担的损失越大,从而有利于行人群体合作行为的涌现;行人密度越大,行人行为越不容易改变,在相同的行人密度和惩罚力度条件下,行人行为概率随时间变化的趋势与行人初始状态无关。最后,基于行人的外部视角,建立复制动态演化模型,研究有管理部门引导者参与的情况下行人选择竞争行为与合作行为的演化规律。在高密度行人紧急疏散过程中,管理部门的引导者采取有效的引导策略对保证行人选择合作行为从而进行安全有序疏散起到至关重要的作用。
[Abstract]:With the continuous improvement of people's living standard and the accelerating process of urbanization, the population and resources of all kinds of public places in modern cities are continuously concentrated, and the threats of various kinds of risks and unconventional emergencies are emerging day by day. The consequence of crowd catastrophic accidents caused by unconventional emergencies during pedestrian evacuation such as crowded stampede and so on is often very serious and the social impact is also very bad. People pay more and more attention to urban safety. In order to ensure the safety of people, it is of great significance to carry out scientific research on the simulation of high-density pedestrian evacuation in public places, risk assessment and emergency management. At present, evacuation research has attracted scholars from different professional backgrounds and become one of the most valuable topics. This paper focuses on the cellular automata model and evolutionary game theory model. The main contents include the following aspects: first, based on the cellular automata model, In the process of pedestrian evacuation, the benefit parameters of competitive behavior and cooperative behavior are constructed, and the rules of pedestrian movement are determined by the value of income parameters, and then simulated under the conditions of different system size, exit width and pedestrian density. Statistic evacuation time. Through comparative analysis, it is found that the pedestrian evacuation time increases linearly under the condition of constant system size and exit width. Under the condition of constant system size and pedestrian density, the pedestrian evacuation time decreases exponentially with the increase of exit width, and the cooperative behavior of high-density pedestrians in emergency evacuation is beneficial to improve the evacuation efficiency. Secondly, in order to urge high-density pedestrians to choose cooperative behavior under emergency evacuation, a cellular automata model of prisoners' dilemma under punishment mechanism is proposed. Using the weak Prisoner's Dilemma to quantify the benefits of pedestrian behavior and using the punishment force as the punishment for the pedestrian's choice of competition behavior, the relationship between the pedestrian's competitive behavior and the degree of punishment in the Von Neumann neighborhood and the Moore neighborhood under different initial conditions is discussed. The simulation results show that the pedestrian density is 0.6 and 0.8 respectively. Through analysis, it is found that the speed of pedestrian behavior is related to the number of neighbors in the evacuation process, and the more the number of neighbors, the less easy it is to change pedestrian behavior. The stronger the punishment, the greater the loss of the pedestrian who choose the competitive behavior, which is beneficial to the emergence of the cooperative behavior of the pedestrian group. The higher the pedestrian density, the harder it is to change the pedestrian behavior. Under the same pedestrian density and punishment intensity, the trend of the change of pedestrian behavior probability with time is independent of the initial state of the pedestrian. Finally, based on the external perspective of pedestrians, a replica dynamic evolution model is established to study the evolution law of pedestrian selection competition behavior and cooperative behavior with the participation of managers. In the process of high-density pedestrian emergency evacuation, the effective guidance strategy adopted by the management plays an important role in ensuring the pedestrian to choose cooperative behavior and carry out safe and orderly evacuation.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP301.1;D63

【参考文献】