Iacppo : A Deep Reinforcement Learning-Based Model for Warehouse Inventory Replenishment

Inventory cost is one of the main costs in Supply Chain Management (SCM), and a reasonable replenishment strategy can reduce the warehouse operation cost. The traditional replenishment strategy is difficult to meet the complex and changing demand of real warehouse scenarios, and because of the spatiotemporal heterogeneity of commodity demand and inventory cost leads to the need to train different commodities one by one using time series prediction models, which greatly increases the modeling cost. To address the above problems, we propose a replenishment model IACPPO that incorporates the A2C algorithm and the PPO algorithm. Firstly, we introduce gated recurrent unit (GRU) into the Actor-Critic network to extract valid information about the environmental state sequences through memory inference; secondly, we use the attention mechanism in the Actor-Critic network to capture the key state sequences; thirdly, we fuse the A2C algorithm with the PPO algorithm to train the whole network simultaneously to obtain the replenishment strategy. Experimental results on inventory datasets show that the replenishment strategy obtained by the IACPPO model achieves lower inventory costs compared to both the seven prediction models and the seven reinforcement learning algorithms. This paper investigated the hyperparameters of the IACPPO model and determined the range of hyperparameter settings such as learning rate and reward scale