시뮬레이션 기반 AGV 층간 물류 시스템의 병목 분석 및 최적 운영 대수 결정에 관한 연구

Abstract

This study proposes a systematic methodology for determining the optimal fleet size of Automated Guided Vehicles (AGVs) in a multi-floor logistics system with spatial constraints, typical of automating existing "brownfield" facilities. Recognizing the complexity of interactions between AGVs and shared resources like elevators, this research utilizes discrete-event simulation to analyze system performance and investment efficiency. A virtual model of a logistics process operating between a basement and a third floor, connected by two elevators, was developed using FlexSim software. The number of AGVs was incrementally increased from one to ten in various scenarios to observe changes in key performance indicators (KPIs), primarily system throughput and resource utilization. The simulation results reveal a clear non-linear relationship between the number of AGVs and system performance. Throughput increased almost linearly as the fleet grew from one to five AGVs, indicating that the AGVs themselves were the primary bottleneck. However, the system reached its peak throughput with five AGVs. Deploying six or more AGVs resulted in diminishing returns, with throughput decreasing due to a bottleneck shift from the AGVs to the fixed-capacity elevators. This shift led to a sharp decline in AGV utilization and increased congestion. Consequently, the study identifies five AGVs as the optimal "elbow point," where throughput is maximized while maintaining high resource utilization and operational stability. The findings provide a practical, data-driven framework for decision-makers to avoid over-investment and achieve a balanced, efficient automated system in complex, multi-floor environments.