Modelling with Neural Networks and Time-Series Forecasting Inventory Control and Cost Reduction in Supply Chain Process

The study delves into the contemporary challenges of production planning, encompassing aspects such as mass customization, fluctuating demand, and intense competition within the supply chain. It underscores the critical need for precise planning while maintaining adaptability to unforeseen disruptions. Recent efforts have centered on leveraging neural networks for real-time processing of extensive, multidimensional datasets to effectively address these challenges. Inventory management plays a pivotal role in mitigating supply chain costs and ensuring organizational success by optimizing inventory levels to minimize shortages, surpluses, and their associated expenses. The study specifically investigates the impact of inventory holding costs on determining overall production costs and timelines. It suggests employing the Auto-Regressive Integrated Moving Average method for forecasting optimal production costs and juxtaposes this approach with a two-layer feedforward neural network. The research aims to bridge gaps in inventory management concerning degraded items and shipment protocols. Its discoveries offer valuable insights into decision-making criteria like product shipment and degradation management, leading to a gradual reduction in total inventory costs. The study’s conclusions are reinforced by sensitivity analyses and numerical evaluations. Future research endeavors should concentrate on enhancing scalability and reproducibility for practical implementation.