突发事件应急管理面临多并发决策任务,相应的决策建模分析要求多种分布式、异构模型的集成。本文针对突发事件应急管理,提出了本体驱动的多模型集成框架。该框架构建了四个基本层,即情景空间层、任务网络层、模型组合层及数据接口层,并采用统一的概念和本体建模语言,分别建立了任务本体、模型本体和数据本体,实现任务知识、异构模型及数据资源的一般化表达与管理。该框架还设计语义推理引擎,驱动了相邻层之间的沟通与合作。通过层级间的交互推理,系统能够语义地分析决策需求,自动选择并组合功能与接口一致的模型,从而实现多模型集成建模。应用示例结果证明了该方法能够为突发事件应急管理的集成建模与仿真提供有效的决策支持。
Since the response of complex emergencies depends on the comprehensive prediction and evaluation, there is increasing attempts to compose multi across-domain models to find details for decision making. In this paper, a four-layer framework is proposed to support the model integration. In this framework, the universal communication between four layers respectively named as Scenario space, Task network, Model composition and Data interface is driven by three engines. With the help of ontology model of task, model and data, the system can semantically search, select and compose models. This framework has capability of supporting model integration dynamically. The results of integrated simulation case indicate that this approach can provide effective support for complex emergency decision making.