基于案例推理的水上交通突发事件应急响应资源需求预测

Prediction of resource for responding waterway transportation emergency based on case-based reasoning

针对水上交通突发事件应急响应的资源需求问题,提出基于案例推理(CBR)的资源需求预测方法。首先,通过分析水上交通突发事件范围和近年来长江干线发生的历史案例,在保证案例完整性和有利于提高检索效率的前提下,提取案例的特征属性,定义案例的结构化表示框架,并讨论水上突发事件应急响应资源需求的特点和内容。然后,根据事件信息的属性类型,定义不同的相似度计算方法,确定案例检索过程中的属性权重,采用最邻近法计算全局相似度,求解资源需求预测结果。案例分析表明,案例2与目标案例的相似度最高,经专家对其资源调用情况进行修正后,作为资源需求预测的结果。

A CBR approach to predicting demand for waterway transportation emergency resources was worked out in this paper. Firstly,through analyzing the scope of waterway transportation emergency events and history cases of maritime accidents in the main channel of Yangzte river,characteristic attributes the cases were extracted,which can both ensure integrity of cases and improve the retrieval efficiency,and frame-based representation of cases was defined. Characteristics of emergency resources demand in waterway transportation emergency were also discussed. And then,a method of calculating the similarity with different attribute values was introduced. Weight values of attributes were determined. The K-nearest neighbor method was adopted to calculate the overall similarity,so as to solve the resource prediction problem. Finally, example analysis demonstrates that resource demand can be predicted by means of this method.