Predictive Analysis of End to End Inventory Management System for Perishable Goods

• Published in: 2022 3rd International Conference for Emerging Technology (INCET) pp. 1 - 5
• Main Authors: Sakhare, Kaustubh Vaman, Kulkarni, Isha
• Format: Conference Proceeding
• Language: English
• Published: IEEE 27.05.2022
• Subjects: Costs; Essential goods; linear programming; Medical services; Oils; Perishable Items; Reinforcement learning; Supply chain management; Supply chains; Transportation
• DOI: 10.1109/INCET54531.2022.9824831

Perishable items are the ones that quickly become spoiled or unsafe for consumption and usage. They are the core of supply chain management due to the limited shelf-life. There are many forms of perishable items viz fresh food products like fruits and vegetables, dairy, poultry, frozen or processed foods, commodities like cosmetics, health care products like medicines, oxygen and lifesaving items like blood. The paper addresses issues of the order management system in the estimation of the demands of the essential perishable items based on its fluctuating demands. The prevalent work was not able to correctly estimate the demands of the items labeled as essential goods. This article describes the identification of the most relevant features to estimate the replenishment policies. The work carried out contemplates the feature importance, exceptionally taking into account the effect of calamities or celebrations and festivities. The novelty of the work lies in comparing the complex two-step models using linear programming and reinforcement learning. The proposed model with simple machine learning algorithms is seen more effective for small scale businesses. The devised model focuses on accurately predicting the order estimate of perishable product under dynamically impacting factors in the supply chain management.