| dc.description.abstract | In a fast-paced and highly regulated pharmaceutical industry, developing an inventory replenishment policy is a critical task due to its unique characteristics, including regulatory compliance, product expiration, and unpredictable demand. In addition, it is highly crucial to quickly adapt the changes in demand in dynamic pharmaceutical market to maintain high service level. The project develops an optimal inventory replenishment policy with Deep Reinforcement Learning (DRL) to ensure the availability of medications while minimizing stockouts and medical waste due to expiration. It relies on continuous learning in which each retailer environment captures the information of dynamic demand patterns, current inventory levels, open orders and lead time as state space to map the inventory problem as Markov Decision Process (MDP). For accurate decision-making in pharmaceutical supply chain, the suitable order quantities are selected from continuous action space which results into higher profitability and serve an increased number of patients, thereby delivering health as a social good in an effective manner. | |
