Research On Demand Forecasting Of Emergency Supplies Based On Case Based Reasoning

In recent years,a varity of natural disasters and sudden public health incidents occur frequently,in order to do post-disaster emergency relief work to improve the efficiency of the rescue,to maximize the reduction of disaster losses.The key issue is to be able to determine the amount of emergency supplies needed for the disaster in the first place so that precious and limited supplies can be delivered to the affected people in a timely manner without causing a greater waste so that each material can be totally used.Do the emergency material demand forecast work can give the relevant departments to actively respond to the implementation of rescue activitity,to provide theoretical basis.Case-based reasoning(CBR)was used,in the dissertation,to solve this problem,and the specific case data can be used as the reference of the current problem.In the decision-making used a lot of image thinking,not simply rely on the rules,but also taking the experience of experts into account.Compared with the traditional model-based approach,CBR does not need to explicitly express the domain model,suitable for solving the problem of structural problems.CBR solves new problems by adjusting the successful solutions of previous issues.To determine the form of expression of the case,the different form of expression of the case have different effects on the CBR process.The ontology method can not only extract the numerical characteristics of the attributes but also extract the feature attributes of the semantic type,moreover describe the details of the case in more detail.The contribution of each feature to the overall similarity is measured by weight,and the weight values of each attribute are determined bu analytic hierarchy process and kernel distance method respectively.And examine their similarities and differences.For the added semantic information,the calculation method based on the conceptual instance not only consider the distance between the concept instances and the upper concept,but also the information at the level of the concept instance,which overcame the deficiency of semantic split calculation in Wu-palmer algorithm.The value of the global similarity is the weighted sum of the numerical similarity and semantic similarity.A simila case obtained by the search will probide a reference for solving the problem case.On this basis,the framework of emergency demand forecast system based on ontology case representation and case representation and case reasoning is constructed.Based on the frame work structure,combined with the field of emergency material needs expertise,the use of Protégé to build emergency materials demand domain ontology.