Optimizing Order Release and Warehouse Forecasting in Eyewear Manufacturing via Linear Programming and Time Series Analysis

This thesis focuses on the development of an automated order release tool and a predictive algorithm to improve the availability of component kit, thus enhancing production efficiency. The solution, based on a real-world case study carried out during an internship in an eyewear manufacturing plant, utilizes data analysis and machine learning techniques, including time series forecasting and linear programming, to provide actionable insights. The thesis also discusses the process of integrating human oversight into automated systems, ensuring daily target achievement and optimizing warehouse inventory levels. Various methods were evaluated and compared to improve the predictive accuracy and operational efficiency, with attention to data preprocessing, model evaluation, and the practical implications of the solution for production planners.