基于CBR的突发事件应急物资动态需求预测方法研究

Dynamic Demand Forecasting Method of Emergency Supplies Based on CBR

针对突发事件应急救援环境动态变化条件下的持续性物资需求问题,在系统分析应急物资需求影响因素的基础上,采用案例推理(CBR)理论与方法,构建了契合突发事件突然性和非例行性,并充分反映适时性、阶段性、随机性和不确定性等需求特征的应急物资动态需求预测模型,给出了面向对象的案例组织与表示模式、相似案例集检索方法、基于灰色关联分析的最佳相似案例匹配算法、以及案例输出与调整方法等,并通过算例进行模型有效性验证。该方法可以根据应急救援进展、灾害演变、物资消耗与筹集等情况,按需预测出下一阶段或某一时段的应急物资种类及数量需求,适用于突发事件应急救援各个阶段应急物资的需求分析,能够有效提高应急救援效率。

For the continuous material needs under dynamic environment conditions of emergency rescue, a dynamic demandforecasting model of emergency supplies, which fits the sudden and non-routine of emergency events, and fully reflects the characteristicsof real-time, stage, randomness and uncertainty of emergency material needs, is constructed on the use of case-based reasoning（CBR）theory and method based on the systematic analysis of emergency supplies demand factors. Furthermore, a model of case organizationand representation, a method of similar cases set retrieving, a matching algorithm of optimal similar case based on grey relational analysis,and a method of case output and adjustment are put forward. At last, the validity of the model is verified by an instance. The dynamicdemand forecasting method can predict the demand of the type and the number of emergency supplies in the next phase or a certain periodaccording to the cases of the emergency rescue progress, the evolution of the disaster, the materials consumption and raise, etc., and isapplicable to emergency supplies needs analysis of all stages of emergency rescue, which can improve the efficiency of emergency rescue.