Forecasting inventory for the state-wide pharmaceutical service of South Australia

Forecasting pharmaceutical stock inventory is a complex problem affecting the healthcare industry across the globe. South Australia is no different. This prompted SA Pharmacy to collaborate with a team of researchers at the University of South Australia to explore ways of projecting inventory for their state-wide services. In this paper, we utilised industry-supplied time-series data to train prediction models using linear regression, exponential smoothing, Holt Winters Seasonal Additive, and Holt Winters Seasonal Additive + damped algorithms. Among these models, the study identified the Holt Winters Seasonal Additive + damped algorithm is the best performing based on an RMSE of 408. We developed a dashboard to consolidate this data and visualise organisational metrics related to optimizing pharmaceutical stock inventory for SA Pharmacy. This study presents our findings and our pharmaceutical dashboard.