Reinforcement Learning Algorithm for One-warehouse Multi-retailer Inventory Problem

We considered a supply chain inventory scheduling problem in which a central warehouse serves n-retailers under. Mathematical model was developed to obtain the optimal revenue for the proposed policy and the objective function is to minimize revenue per unit time. Reinforcement learning algorithm was used to solve the mathematical model, including state determination criteria, reward function, and other parameters. After determining the appropriate parameters of the algorithm, the effectiveness of the algorithm was verified by numerical example. The performance of the algorithm was tested for randomly generated problems. Results show the robust performances of the proposed algorithm.