A data driven inventory solution framework for an inventory management problem

Official URL: https://risc01.sabanciuniv.edu/record=b2756179_(Table of contents)

Demand forecast is the most essential input of the inventory models. In the case of manufacturing processes with a variety of similar products that can use a shared production line and common resources, the total amount of inventory and the itemized inventory levels need to be determined separately, but considering the correlation caused by the shared resources, we propose a framework that calculates the total required inventory levels based on the previous sales and demand forecasts and then determines the maximum amount of a production to be inventoried as a function of each product’s forecast, and its previous sales for the period of the inventory. After deriving the max ratio to produce for each product, we propose clustering the products based on this ratio, to facilitate the application in industry. Using these ratios and the forecasts, the amount that need to be produced for each product is calculated. Then a new ratio for each product is calculated by dividing the amount of product to the required inventory for that product. Then the extra capacity is used so lowest ratio will become as high as possible. In our case study, we applied the framework to a tire cord fabric manufacturer (Company K), and after implementation they reported a total inventory decrease from 20 days of service to 10.