Reordering Point Determination Using Machine Learning Technique for Inventory Management

Reordering point is an important aspect of inventory management. An optimal determination gives benefit in safety stock management, increase productivity, reduce inventory costs, and increase revenue. Traditional methods based on mathematical functions have achieved a certain degree of success. However, it is time consuming for inventory officers to calculate the reordering points for all products and update the value periodically. It is not adaptable to other factors that are equally significant but not a part of the mathematical functions. This paper attempts to use a machine learning technique with an artificial neural network model, to determine the optimal reordering points. An artificial neural network model was developed using MATLAB R2016a. The data used for training was collected from companies of various industries and was made publicly available online. Different training algorithms were applied to the artificial neural network model. An artificial neural network with the best performance is identified with a MSE in the order of 10-5 and an adjusted R2 value of 0.99999. This result can be used as a basis for further development of an inventory management module of an open source ERP program.