Transforming Supply Chain Operations through AI and Machine Learning: Real-Time Demand Forecasting, Inventory Optimization and Logistic Systems

Kamana Parvej Mishu; Ishfaq Ur Rahman; Mohammad Tahmid Ahmed; Apurbaa Sarker; Afia Masuda Supti; Nadira Kulsum Papri

doi:10.63332/joph.v6i3.4059

Authors

Kamana Parvej Mishu College of Graduate and Professional Studies, Trine University, Angola, IN 46703, USA
Ishfaq Ur Rahman College of Graduate and Professional Studies, Trine University, Angola, IN 46703, USA
Mohammad Tahmid Ahmed College of Graduate and Professional Studies, Trine University, Angola, IN 46703, USA
Apurbaa Sarker Department of Graduate Information Technology, University of the Cumberlands, Williamsburg, KY 40769, USA
Afia Masuda Supti Department of Marketing Analytics and Insights
Nadira Kulsum Papri Department of Graduate Information Technology, University of the Cumberlands, Williamsburg, KY 40769, USA

DOI:

https://doi.org/10.63332/joph.v6i3.4059

Keywords:

Demand Forecasting, Inventory Management, Machine Learning, Deep Learning, Reinforcement Learning, Supplier Collaboration, Supply Chain Optimization

Abstract

The optimization of the supply chain is currently becoming more and more important in the context of businesses that are interested in improving their operations in terms of efficiency and flexibility in a dynamically changing global context. The incorporation of Artificial Intelligence (AI) and Machine Learning (ML) within the sphere of supply chain businesses offers the next-generation possibilities of predictive analytics, demand forecasting, and inventory optimization, as well as decision-making, which can be viewed as a disruptive opportunity of the contemporary businesses. The proposed research is expected to examine the ways in which AI and ML technologies should be utilized to enhance demand forecasting, inventory optimization, and a logistics system in supply chain management. The study is aimed at ensuring the creation of an AI-ML model that is more accurate in terms of its predictions, real-time adjustments in inventory, and dynamic cooperation with suppliers. The hybrid AI-based approach is suggested, which consists of Recurrent Neural Networks (RNNs), Attention Mechanisms, and Ensemble Learning forecasting. Inventory optimization based on Deep Learning (DNNs) and dynamic supplier collaboration framework are also considered important elements. The USA retail supply chain dataset data are collected, preprocessed and it is used to train and evaluate the models. The findings show the vast improvement of the accuracy of demand forecasting with the hybrid AI model reducing the forecast error by half and a 21.4% increase in the efficiency of the inventory. The real-time optimization strategy decreased stockouts by a great margin and minimized the holding cost. Moreover, the supplier welfare system boosted the risk reduction measures and increased the resilience of the whole supply chain. The data set to be utilized consists of transactional data, sales data, customer behavior, and external market information, which are obtained within the Retail Supply Chain Sales Dataset of Kaggle.