ASRS, AMR integration with WMS - challenges and solutions

Purpose- The study investigates the integration of Automated Storage and Retrieval Systems (ASRS) and Autonomous Mobile Robots (AMRs) with Warehouse Management Systems (WMS), aiming to enhance operational efficiency through machine learning. Key challenges addressed include data compatibility, real-time decision-making, and effective resource allocation. Methodology- Machine learning models were applied to optimise system performance: Bayesian Neural Networks (BNNs) for demand forecasting, Random Forests for resource allocation, K-means clustering for task prioritisation, and Support Vector Regressor (SVR) for performance evaluation using Mean Squared Error (MSE). Findings- BNNs improved demand prediction, enabling adaptive adjustments of ASRS and AMRs. Random Forests efficiently optimised resource distribution, while K-means clustering successfully prioritised high-demand tasks to support lean operations. The SVR achieved an MSE of 2.47, confirming low prediction error and model effectiveness. Conclusion- Integrating machine learning into ASRS-AMR-WMS systems provides a scalable framework for modern warehouses, fostering real-time adaptability, improved resource utilisation, and enhanced productivity. Keywords: ASRS, AMR, WMS integration, machine learning, Bayesian Neural Networks, Random Forest, K-means clustering, demand forecasting, resource optimisation, warehousing efficiency