面向案例消耗推理的应急物资预测

Forecast of Emergency Supplies for Case Consumption Reasoning

面向案例推理(Case-Based Reasoning,CBR)的应急物资需求预测中,针对老旧案例影响推理结果精确度的问题进行了研究,给出了一种基于消耗策略的案例推理的应急物资预测方法。通过粗糙集属性依赖度的计算确定了案例属性之间的权重系数;针对地震应急数据特征提出了一种消耗策略的定义,确定消耗函数、消耗参数和消耗区间参数,采用消耗策略对各案例间的相似度进行优化调整,减小老旧案例的权重,进而不同程度地削弱老旧案例对案例匹配结果的影响,再从案例库中检索到与目标案例匹配的最佳源案例,从而决策出目标案例的处理方案;通过进行地震实例分析,验证了参数调节后的案例消耗推理的预测结果精度更高。该方法在应急救援的物资预测中有一定的借鉴意义。

In the Case-Based Reasoning(CBR)of emergency supplies demand forecasting,aiming at the problem that the old cases affect the accuracy of reasoning results,this paper proposes a method of emergency supplies forecasting based on a case-based reasoning with consumption strategy.Firstly,through the calculation of attribute dependency of rough set,the weight coefficient between case attributes is determined.Secondly,according to the characteristics of earth-quake emergency data,a definition of consumption strategy is proposed,the consumption function,consumption parame-ters and consumption interval parameters are determined,the consumption strategy is used to optimize and adjust the simi-larity of each cases,weaken the weight of old cases,weaken the influence of old cases on case matching results in varying degrees,and then retrieve the best matching with the target case from the case database and get the processing scheme of the target case.Finally,through the analysis of an earthquake case,it is proved that the prediction result accuracy of case consumption inference after parameter adjustment is higher.Therefore,this method can be used for reference in the sup-plies forecast of emergency rescue.