基于进化博弈和心理学理论的群体暴恐事件仿真与推演

发布时间：2018-01-15 09:25

本文关键词：基于进化博弈和心理学理论的群体暴恐事件仿真与推演　出处：《郑州大学》2017年硕士论文　论文类型：学位论文

【摘要】：群体暴恐事件的发生率近年来在全世界范围内呈上升趋势,严重威胁世界和平,并对人类的公共安全造成危害,众多群体暴恐事件的惨痛案例所带来的警示教训,促使我们有必要深入的研究群体暴恐管控的仿真推演技术。群体暴恐管控过程本质上可被视为由心理机制驱动的进化博弈过程。因此,本文提出了一种基于进化博弈和心理学理论的群体暴恐事件人群行为的仿真方法。该方法综合考虑在群体暴恐事件中人群的情绪感染机制与策略的进化博弈机制,对群体暴恐事件的复杂群体行为进行仿真推演。具体地,我们采用多智能体技术建立群体暴恐的基本计算框架,在虚拟人工社会中模拟暴恐场景和各种不同类型的虚拟智能体群体,设定智能体的基本属性,并根据智能体类型的不同制定相应的行为交互规则。群体的情绪变化会影响其应激运动过程。基于此,我们对群体暴恐事件中的智能体的情绪感染机制进行分层处理,综合考虑智能体受到感知范围内其他智能体所感染的外部情绪与博弈交互后智能体自身所产生的内部情绪激励,对群体暴恐事件中智能体的情绪感染进行更新计算。而后,针对群体暴恐环境中“警察——暴徒”之间明显的博弈关系,采用进化博弈理论对智能体的策略演化更新,从而实现“环境——感知——情绪——策略——行为”的群体暴恐行为映射流程,较好的推演群体暴恐事件的发展态势。最后,本文通过仿真实验,模拟了在不同初始条件下群体暴恐事件中的人群变化态势,分析了智能体的情绪值与群体暴恐发展态势的关系,推演了智能体在群体暴恐中策略所作出的决策机制,分析了警察与暴徒之间胜负的条件。实验表明,基于进化博弈和心理学理论的群体暴恐事件人群行为的仿真方法可以呈现出群体暴恐事件的宏观态势,并在微观上刻画出个体间情绪的感染机制和策略的进化机制,为全面认识暴恐事件演化过程进而预知事件演进态势并主动应对提供了理论支持。
[Abstract]:The incidence of mass violence has increased worldwide in recent years, posing a serious threat to world peace and endangering human public safety. Warning lessons from the tragic cases of mass terror. It is necessary for us to study the simulation and deduction technology of group riot control. In essence, this process can be regarded as an evolutionary game process driven by psychological mechanism. In this paper, a simulation method of crowd behavior of group violent and terrorist events based on evolutionary game and psychological theory is proposed. This method synthetically considers the mechanism of emotional infection and the evolutionary game mechanism of strategy. . The complex group behavior of mass violence events is simulated and deduced. In particular, we use multi-agent technology to establish the basic computing framework of mass violence. In the virtual artificial society, the scene of violent terror and different types of virtual agents are simulated, and the basic properties of the agents are set. And according to the different types of agents to formulate the corresponding rules of interaction. The emotional changes of the group will affect their stress and movement process. Based on this. We delaminate the emotional infection mechanism of agents in mass violence and terror events. Considering the external emotion of the agent infected by other agents in the range of perception and the internal emotional stimulation generated by the agent itself after the interaction between the agent and the game. The emotional infection of agents in mass violence and terror events was updated and calculated. Then, the obvious game relationship between "police and mob" in the environment of mass violence and terror was analyzed. The evolutionary game theory is used to update the strategy evolution of the agent, so as to realize the mapping process of "environment-perception-emotion-strategy-behavior". Finally, this paper simulates the crowd change situation in different initial conditions through simulation experiments. This paper analyzes the relationship between the emotional value of the agent and the trend of the development of group violence, deduces the decision-making mechanism of the agent's strategy in the mass violence and terror, and analyzes the conditions of victory and defeat between the police and the mob. Based on evolutionary game theory and psychological theory, the simulation method of crowd behavior of group violent events can present the macroscopic situation of group violent and terrorist events. At the same time, the infection mechanism of emotion and the evolutionary mechanism of strategy among individuals are described in the micro level, which provides theoretical support for the comprehensive understanding of the evolution process of violent and terrorist events and the prediction of event evolution situation and active response.
【学位授予单位】：郑州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：B842;D035.3

【相似文献】