基于基因结构的复杂应急案例表示方法

案例推理作为人工智能领域类比推理的有效方法,其检索效果与推理能力很大程度上依赖于案例构建的合理与否.考虑到已有方法在表示复杂应急案例上的不足,本文依据历史案例将复杂突发事件的"情境-情景"结构抽象化,以此提出复杂应急案例基因模型的概念.基于基因模型构建多级的案例结构,能形成可重复使用的、标准的、简化的、可供应急决策者重组参考且具有"生物型"特征的复杂应急案例表示模型.该方法在一定程度上继承与集成了框架与本体表示法的特点,实现了复杂应急案例结构的系统性表达,使案例内容有机连接成整体,可更好地服务于复杂应急案例推理.为更好地说明,本文以"台风/暴雨"引发的复杂情境链为例,阐述基因结构表示复杂应急案例的合理性与有效性.

Complex emergency case representation method based on genetic structure

Case-based reasoning (CBR) is an effective method of analogism in field of artificial intelligence. The retrieval effect and reasoning capacity of CBR depend, to a great extent, on whether the case building is reasonable or not. Regarding the shortages of the existing methods for representing complex emergency case, in this paper the concept of complex emergency case gene model is proposed by abstracting "context-scenario" structure of complex emergency on the basis of historical cases. The multi-level case structure is built based on gene model, which can form a reusable, standard, simplified complex emergency case representation model with "biological type" characteristics for emergency decision-makers to refer. This method inherits and integrates the features of Frame and Ontology representation method to some extent. Also, it realizes to express complex emergency case structure systematically and connect case content into a whole, which can be better to serve complex emergency CBR. For better explanation, in this paper a complex context chain caused by "typhoon/rainstorm" is taken as an example to illustrate the reasonability and effectiveness of genetic structure for representing complex emergency case.