Deep Reinforcement Learning in Inventory Management
"IoT-driven dynamic replenishment of fresh produce in the presence of seasonal variations: A deep reinforcement learning approach using reward shaping", recently published in Omega
The Internet of Things (IoT) is reshaping inventory management through gathering real-time data across the supply chain, which is particularly valuable in the fresh produce industry. Due to short shelf life and seasonal fluctuations, fresh produce suppliers struggle with when and how much to restock. This paper introduces DQN-SV-RS, a deep reinforcement learning approach that utilizes IoT data to make better decisions about ordering and stocking. A key feature of DQN-SV-RS is “reward shaping,” which helps adapt to seasonality, reduces produce waste, and ramps up sales. DQN-SV-RS is a promising, scalable model for managing perishable goods effectively.
By Zihao Wang, Wenlong Wang, Tianjun Liu, Ai-Chih Jasmine Chang, Junmin Jim Shi.