基于CBR的重大环境事故应急响应专家系统研究进展与展望

Research progress and prospect review over the CBR-based expert system for major environmental accident emergency response

针对当前社会环境事故频发现象,分析国内外重大环境事故应急响应策略的研究进展,指出目前应急响应决策存在的问题。分析了基于案例推理(CBR)的理论和应用研究现状,拟将CBR技术引入重大环境事故应急响应领域,提出了基于案例推理的重大环境事故应急响应专家系统(CBR-MEAERES),并介绍了CBR-MEAERES的结构和核心问题。

The paper is aimed to make a research advance and prospective review over the CBR-based expert system fur major environmental emergency response. While giving a comprehensive analysis over the current research situations on the emergency response strategies to the major environmental accidents, we have pointed out the shortcomings or deficiencies of the research in this way. We all know that, in recent years, serious environmental accidents or disasters have been taking place in our country quite frequently, which has become an essential factor and brings threats and destructive force to the sustainable development of our regular eco-envirenment. There-fore, it becomes imperative to take immediate rescue operations to save our environment. For this purpose, we have studied the casebased reasoning （CBR）, major environmental accidents, some emergency response measures with reference to the existing expert systems. In doing so, this paper suggests that it is imperative to establish a major environmental emergency accidents response expert system in the whole country. And, in turn, we have put forward some theoretical and applied research measures based on the case-based reasoning （CBR） technology, which should be introduced into the system of major environmental acCident emergency response and the framewOrk of CBR project. Furthermore, the framework and the core issues of CBR-MEAERES have also to be briefly introduced into the expert response system, including man-machine interface, case retrieval engines, case base and some other parts. The other factors or indexes of the system should also be included, such as the accident time, accident location, accident incentives, accident types, the pollutant type, pollutant emissions and distribution of the receptors as feature attributes, in accordance with the predefined similarity computation formulae, with the CBR-MEAERES being used to retrieve the most similar cases in the so-callod nearest neighbor method. Moreover, the revised case should be automatically added to the ease base by using the self-taught CBR-MEAERES. And, then, when the base base gets enough expanded, CBR, MEAERES will be able to screen out quickly and accurately the emergency response strategies to major environmental accidents, and then provide sufficient guiding significance to the emergency response of the major environmental accidents.