Applying Machine Learning For Demand Forecasting And Linear Programming For Aggregate Production Planning

Abstract

Demand forecasting is, without a doubt, the most crucial aspect of any company's supply chain. Especially, the new century of big data gathered from complicated customer behaviors, as well as severe competition among businesses in the same areas, has prompted many organizations to invest in improving forecasting precision. After demand forecasting, aggregate production planning is concerned with determining production, inventory, and labor levels in order to meet fluctuating demand requirements over a six-month to one-year planning horizon. The aim of this thesis was to design a machine learning model to propose sales prediction and aggregate production planning for AJF Company – a multinational company that manufactures and sells their main product is Korean food. The research also includes procedures for improving model output, such as parameter adjustment, feature selection, and the removal of outlier values. The findings show that the mistake rate and processing time are both respectable, with plenty of potential for improvement. From that, AW could improve the demand forecasting accuracy, which leads to better performance for the whole supply chain of the company