基于ACP行为动力学的犯罪主体行为平行建模分析

Parallel Modeling of Criminal Subjects Behavior Based on ACP Behavioral Dynamics

犯罪行为分析是侦查破案的重要参考,也是学界长期以来关注的热点.目前,犯罪行为分析主要采用现场证据-行为推断的思路,忽略了犯罪过程中犯罪主体和客体之间的复杂互动.本文在基于ACP(Artificial societies(人工社会)+Computational experiments(计算实验)+Parallel execution(平行执行))方法的犯罪现场、平行系统框架下,从行为动力学角度提出了故意杀人行为的犯罪主体时间和空间互动模型,并采用真实案例数据对模型参数进行了标定.计算实验结果表明,本文提出的互动模型能较好地模拟真实数据,从而为分析犯罪过程中的复杂互动提供了一个可靠的基础.

Being important for criminal investigation, criminal behavioral analysis is always an active research topic among scholars. Current methods of criminal behavioral analysis mainly adopt the approach of evidence-behavior inference, and ignore the complex interactions between the criminal and the victim. In the framework of ACP （artificial societies computational experiments ＋ parallel execution）-based parallel system, this paper proposes temporal-spatial interaction models for artificial criminals from the behavioral dynamics point of view. In addition, actual intentional homicide data is used to demarcate the parameters of the models. Results of computational experiments indicate that the proposed interaction models can simulate the criminal frequencies very well according to the actual data source. Thus they provide a solid foundation for the complex interactions in criminal behavioral analysis.