Inventory Management System Using Reinforcement Learning: A Case Study

This research intends to put RL methods related to SCM to use in the management of input stocks.Estimating the composition of a small retailer's inventory system, specifically to recharge Coke sales, the research aims to improve the forecast of merchandise, when they should be refilled, to fulfill client expectations.The deep Q network (DQN) algorithm is used to represent the objective of the study comparing the performance of the RL-based inventory control strategy with the classic static control method ((s, S) inventory control) in a numerical test.These financial parameters are determined along with other operational constraints, such as inventory capacity, lead time, and product order costs.The demand patterns between weekdays and weekends form the basis for the simulation of historical desire data to train DQN model.The comparison of RL-based methods in the retail industry supply chain is covered by this study monetarily.Consequentially, the study introduces RLbased methods as one of the techniques in the area of improvement of retail inventory management practical applications with real-life supply chain examples to complement and prove their success.