A machine learning model with linear and quadratic regression for designing pharmaceutical supply chains with soft time windows and perishable products

Abstract

Product perishability is an important problem in Pharmaceutical Supply Chains (PSC). Demand perishability and unpredictability add enormous complexity to successfully implementing efficient Pharmaceutical Supply Chain Network Design (PSCND). This study develops a mathematical model for the PSCND. The two objective functions used in the model minimize the total cost and the delivery penalty due to schedule violations involving Soft Time Windows (STWs). Using a STW strategy in the Distribution Center (DC) allocation model empowers decision-makers to include a time-based performance metric for DC evaluation based on the degree of urgency or need for a part. Moreover, Linear Regression (LR) and Quadratic Regression (QR) Machine Learning (ML) algorithms are proposed to forecast the demand and decrease the possibility of a shortage in the PSCND. We show that QR has better performance than LR in PSCND. In the proposed approach, the demand for medicine is forecasted by the QR technique. The Goal Attainment (GA) method is used to solve the suggested model. Finally, sensitivity analysis and managerial perspectives are offered. The numerical outcomes show that the suggested model leads to an efficient PSC, reducing cost, decreasing shortage, and increasing customer satisfaction with STW consideration.