Role of technology in implementing lean in warehouse operations

Purpose- The study explores the role of machine learning and reinforcement learning in optimising lean warehousing practices, focusing on demand forecasting, inventory optimisation, and stock prioritisation. Lean warehousing aims to reduce waste, cut costs, and maintain efficient stock levels through data-driven strategies. Methodology- Three models were applied: Long Short-Term Memory (LSTM) for demand forecasting, reinforcement learning (RL) with placeholder costs for dynamic inventory management, and K-means clustering for inventory prioritisation. Performance metrics included RMSE, MAE, total reward, and Silhouette Score to evaluate effectiveness. Findings- The LSTM model produced accurate demand forecasts with low RMSE (0.048) and MAE (0.025), aligning stock levels with actual demand. RL recorded a negative reward of −1511.83, highlighting the importance of integrating real-time cost data for better inventory decisions. K-means achieved a strong Silhouette Score (0.935), effectively supporting ABC inventory classification. Conclusion- The study demonstrates that machine learning and reinforcement learning can significantly enhance lean warehousing by improving demand alignment, inventory prioritisation, and operational efficiency.