基于Agent的群体性暴力事件的建模与仿真

对群体环境下个体的犯罪行为选择进行了研究。提出了一种新的犯罪风险评估模型,通过对个体的不满度、冒险等级等因素的简化提取,再现了暴乱发展扩散的情形。仿真实验表明,警察的出现能够有效抑制暴乱时"破窗效应"的扩散,而人群中核心分子和易转化分子对暴乱的规模增长和扩散起到了重要作用。研究结果可以为公安机关处置群体性暴力事件提供重要理论依据。

Agent-based modeling and simulations of collective violent events

Studying the criminal behavior choice of individuals in crowd context.By Abstracting individual’s grievance,risk level and other important factors,we present a new criminal risk assessment model to represent the events’ growing and spreading in virtual world.Considerable simulations indicate that the appearance of police can effectively prevent "windows-breaking effect" spreading in events and the proportion of hard-core and hangers-on plays an important role in the growing and spreading of events.The results provide meaningful theoretical basis for public security organs to deal with the collective violent events.