Deep Learning and Optimization Approaches in IoT-Based Smart Pharmacies for Optimizing Stock Management with Siamese Heterogeneous Convolutional Neural Networks: A Review

Abstract

The rapid growth of digital healthcare technologies has encouraged the integration of Internet of Things (IoT), artificial intelligence, and deep learning techniques into pharmaceutical supply chain management. Smart pharmacy systems utilize IoT-enabled sensors, RFID tags, cloud platforms, and predictive analytics to improve inventory monitoring, reduce drug wastage, and ensure timely medicine availability. Traditional pharmacy inventory systems often depend on manual operations that may result in inaccurate demand forecasting, stock shortages, overstocking, and increased operational costs. IoT technologies provide real-time visibility into inventory levels and support automated stock replenishment, expiry monitoring, and supply chain transparency. Deep learning approaches, particularly convolutional neural networks and Siamese heterogeneous convolutional neural network (SHCNN) architectures, have shown strong potential for optimizing stock management and predicting drug demand patterns. These models can efficiently analyze multimodal healthcare datasets, including pharmacy sales records, environmental monitoring data, and smart shelf sensor information. The integration of machine learning with IoT platforms enhances predictive analytics, operational efficiency, and automated decision-making. This review highlights recent developments in IoT-based smart pharmacy systems and deep learning optimization techniques for pharmaceutical inventory management. The study concludes that combining IoT technologies with SHCNN models can significantly improve inventory accuracy, reduce stockouts, and enhance the efficiency of smart pharmacy systems.