Інтегроване виробниче планування на основі предиктивної та прескриптивної аналітики

Abstract

The paper investigates an integrated approach to decision support in production planning that combines predictive analytics (for demand forecasting) and prescriptive analytics (for resource optimization). A hybrid forecasting model based on SARIMA and Random Forest algorithms with automatic selection of the most effective one has been implemented. To optimize the production plan, a linear programming model aimed at minimizing the total costs of the enterprise has been developed. A web application has been created, and its testing on real data confirmed a 14.5% reduction in total costs and achieved forecasting accuracy with an error of less than 9%.