Study on the guarantee of emergency logistics comprehensive performance based on Neural Network

Emergency logistics involves dispatching commodities to affected areas as soon as possible so that relief operations are accelerated. To strengthen the comprehensive performance management, this article developes back-propagation neural network model to evaluate the comprehensive performance of emergency logistics using five critical variables of cost, inventory volume, accuracy of delivery, quick response, and superiority of facilities as inputs, and comprehensive service level as output with sigmoid function. The neural network programming uses the mainstream Java language design, inserts to the HTML, page and adopts browsers to run the procedure. Data cleated with normalized processing is from a certain military participating in emergency logistics. Through the simulation, it puts deep insights into the importance of different factors to guarantee the emergency logistics performace.