Efficient inventory handling and pallet management play a crucial role in reducing logistics costs and ensuring smooth supply chain operations. This research focuses on challenges such as poor pallet stacking, uneven weight distribution, and inefficient utilization of warehouse space. To address these issues, a machine learning–based framework is proposed for optimizing both load balance and spatial allocation within warehouses. Unlike conventional methods—such as rule-based systems, 3D scanners, and industrial weighing scales—that are often costly and rigid, the proposed approach leverages intelligent algorithms adaptable to dynamic storage environments. The study integrates data from weight sensors and 3D imaging tools, applying advanced optimization techniques including Reinforcement Learning (RL), Linear Programming (LP), and Genetic Algorithms (GA). By combining historical datasets with real-time sensor inputs, the system adapts automatically to varying inventory volumes and box dimensions. Experimental analysis demonstrates that the proposed model improves pallet stacking efficiency and minimizes wasted space, yielding a 15% increase in warehouse throughput and a 10% reduction in pallet waste. These results highlight the potential of data-driven optimization to lower operating expenses, reduce risks of overloading, and streamline pallet movement. Beyond immediate improvements, the study emphasizes the broader implications of machine learning in warehouse logistics—enhancing inventory accuracy, supporting predictive analytics, and paving the way for future developments such as demand-responsive storage allocation, automated load distribution, and live pallet tracking. The findings confirm that integrating intelligent algorithms with real-time analytics can deliver scalable and adaptive solutions for modern warehouse optimization.
Equbal et al. (Thu,) studied this question.