面向应急管理计算实验的模型构建和模型管理

Model development and management in the computational experiment oriented to emergency management

基于人工社会的计算实验使应急管理的定量分析与研究成为可能。然而,应急管理模型涉及多个学科与领域,具有多层次、非线性和多粒度等特点,需要一套规范的流程来指导人工社会的模型建立和模型管理。论文梳理了面向应急管理的模型体系,提出了基于领域特定建模的人工社会建模方法,解决了多领域建模问题;研究了模型的形式化描述与编码问题,并实现了模型的有效管理;结合"甲型H1N1"病程模型,详细论述了模型的构建方法与管理问题,为面向应急管理的计算实验提供一体化的模型服务。

Computational experiments based on artificial society make it possible to study and quantitatively analyze the emergency management process. However, emergency management refers to multi-disciplines and multi-domain, and models have the characteristics of multi- hierarchy, nonlinearity and multi-granularity, so a standard workflow is needed to guide the modeling development and management of artificial society. The model architecture of emergency management was carded, and the artificial society modeling method based on domain specific modeling was introduced to solve the multi-domain modeling problems. Moreover, the formal description methods as well as model encoding were implemented, which promoted the effective management of models. The development methods and management problems were discussed in details in the case study of H1 N1 models, which provided the integrative model services for computational experiments oriented to emergency management.