Convolutional Neural Networks (CNN), which are composed of multiple processing layers to learn the representations of data with multiple abstract levels, are the most successful machine learning models in recent years. However, these models can have millions of parameters and many layers, which are difficult to train, and sometimes several days or weeks are required to tune the parameters. Within this paper we present usage of a trained deep convolutional neural network model to extract the features of the images, and then, used the AdaBoost algorithm to assemble the Softmax classifiers into recognizable images. This method resulted in a 3% increase of accuracy of trained CNN models, and dramatically reduced the retraining time cost, thus, it has good application prospects.

Drug inventory management is an important part of hospital management. The large amounts of drug data in hospitals bring challenges to optimizing the setting values for the safety stock and the maximum inventory of each drug. This study combined a two-stage clustering method with an inventory policy (s, S) and established a simulation optimization model for the case hospital’s outpatient pharmacy. This research used the simulation optimization software Arena OptQuest, developed by Rockwell Automation Inc (Rockwell Automation, Coraopolis, PA, USA), in order to determine the minimum and maximum values (s, S) of the best stock amounts for each drug under the considerations of cost and related inventory constraints. The research results showed that the minimum and maximum inventory settings for each drug in the simulation model were better than those set by the case outpatient pharmacy system. The average inventory cost was reduced by 55%, while the average inventory volume was reduced by 68%. The proposed method can improve management efficiency and inventory costs of hospital pharmacies without affecting patient services and increasing the inventory turnover rate of the drugs.