Simulation-based design optimization of deep-lane autonomous vehicle storage and retrieval systems (AVS/RS) for efficient and sustainable automated picking

AVS/RS robot-to-parts picking systems are an emerging technology in automated warehouses, aimed at improving worker comfort and safety by alleviating the physical strain and fatigue associated with manual picking tasks. Despite their potential, the study of these systems is still highly unexplored due to their novelty. This research addresses this gap by optimizing the operational efficiency and environmental sustainability of these advanced storage systems. We conducted 256 simulations across 32 warehouse configurations, varying the number of bays, tiers, and vehicle parameters, under eight distinct demand scenarios with differing order inter-arrival times, order sizes, and order variability. Using a multi-criteria optimization model solved with the Non-dominated Sorting Genetic Algorithm (NSGA-II), we identified optimal solutions. Pareto curve analysis revealed configurations that balance operational efficiency and environmental sustainability. The results demonstrate that specific configurations significantly outperform others in terms of throughput, service time, space utilization, and energy efficiency, providing a blueprint for designing efficient and sustainable automated warehouses. This work significantly contributes to the development of AVS/RS robot-to-parts picking system, enhancing worker comfort and safety, and advancing the management of warehouse operations.