Research On Performance Evaluation Of Emergency Logistics Based On BP Artificial Neural Network

In recent years,all kinds of unexpected events occur frequently in China,such as the snow and ice disaster in 2008 and the Wenchuan earthquake,the bird flu in 2009 and the flood disaster in the year of 2016.The impact of such emergencies should not be underestimated.They not only resulted in casualties and property losse,but also affected the social stability and development,and even endangered national security.Therefore,emergency management has caused the government,institutions,enterprises and individuals concerned.The key link of emergency management is emergency logistics which plays an important role both in the relief process,such as control of the disaster,protecting the affected people and carrying out disaster relief activities,and posting disaster reconstruction.Combining the comprehensive analysis of the related research of emergency logistics and emergency logistics performance at home and abroad with the characteristics of emergency logistics,and analyzing the influencing factors from the emergency logistics performance at the same time,according to the principles of constructing index system,this thesis put forward the evaluation index system of the new emergency logistics performance.Secondly,it analyzed the advantages and disadvantages of various methods of performance evaluation and established the performance evaluation model based on BP artificial neural network.Finally,The application of the emergency logistics case was analyzed,and the validity and operability of the model were verified.The main research work and conclusions are as follows:(1)The theory of emergency logistics and emergency logistics performance were systematically analyzed.This thesis expounded the characteristics of emergency logistics,the operation stage and operation process on the basis of the concept of emergency logistics.The emergency logistics performance theory was discussed,and the emergency logistics performance factors were deeply analyzed.All of these laid a theoretical foundation for the construction of emergency logistics performance evaluation index system.(2)Combining the index design principles with the characteristics of emergency logistics,the performance evaluation index system of emergency logistics was established.Taking the "four unity" principles of index design and factors of emergency logistics,this thesis built a performance evaluation index system which include economic performance and capability performance of emergency logistics,a total of 25 indicators.(3)The advantages and disadvantages of BP artificial neural network and other performance evaluation methods in emergency logistics performance evaluation were analyzed.Compared BP artificial neural network with other methods,it not only fit for modeling complex problems,but also avoid the interference of subjective factors on the evaluation results.What's more,BP artificial neural network can parallel computing simply,and it has strong robustness and fault tolerance.According to the characteristics of the emergency logistics performance evaluation,such as complex factors and incomplete information,the BP artificial neural network can get more accurate performance evaluation results than other methods.(4)The thesis built the emergency logistics performance evaluation model and empirical analysis.The performance evaluation model of emergency logistics based on BP artificial neural networks was established by training on BP artificial neural network with emergency logistics case data.Taking the flood disaster in Hubei province in 2016 as an example,it was put into empirical analysis by using trained BP artificial neural network model,and then the specific value of emergency logistics performance in the flood disaster was obtained.The above event verified the operability and accuracy of the model.The built emergency logistics performance evaluation model provided new ideas and methods for the evaluation of emergency logistics performance.