Improving Order Fulfillment in Pharmaceutical Supply Chains Using Deep Reinforcement Learning

Abregu-García D.; Camacho-Rondinel N.; Romero-Varona L. · Proceedings of the 2025 IEEE 32nd International Conference on Electronics, Electrical Engineering and Computing, INTERCON 2025 (2025) · DOI: 10.1109/intercon67304.2025.11244663

Resumen

Logistics efficiency is a critical factor for the competitiveness of organizations. Its importance is even greater in sectors such as pharmaceuticals, where unmet demand in the supply chain leads to economic losses for companies and interruptions in medical treatments that harm public health. This study addresses the logistical limitations affecting the efficient fulfillment of orders in a pharmaceutical wholesaler in Peru, which recorded a Fill Rate of only 67% due to recurring stockouts. In response to this situation, the implementation of an intelligent procurement model based on Deep Reinforcement Learning (DRL) is proposed, using the Proximal Policy Optimization (PPO) algorithm. The model identifies optimal purchasing policies by considering inventory constraints, transportation lead times, and uncertain demand. This proposed solution was validated through simulations, which showed an improvement in the Fill Rate to 94% and a reduction in economic losses exceeding S/60,000 per year.

Instituciones: Universidad Peruana de Ciencias Aplicadas