Machine Learning-Aided Component Inventory Reduction in Assemblies

Kwa Shu Ting

Abstract

Inventory management is relatively a concern in supply chain and business especially it directly reflects the inventory holding cost in the business financial report. Inventory can be converted into profit for the business if it can sell away to the customer. Especially for assembly, it requires components to be kept in storage and ready to be assembled once the order is released. Assembly may use a few alternative components, and the components were procured on a per-order basis, those that are not used for the assembly result in excess inventory and hence increase in cost. In this report, I will retrieve company data and use machine learning to help on analysis component inventory in assembly. It will further help to decide whether to eliminate the component, not in use and possibly to choose a similar component as an alternative to eliminate more components. Data were mostly retrieved from SAP and Microsoft Access databases that were used in the company. Microsoft Visual Basic for application will be used to run the data to know how many components are shared with other assemblies. Juypter Notebook or Spyder software will be used as a machine learning tool in this project to run the database and provide the result for analysis.