Simulation-based Deadlock Avoidance and Optimization in Bidirectional AGVS

Abstract

Autonomous guided vehicle systems (AGVSs) are popular in production systems, container ports, and intra-logistics. An efficient deadlock avoidance algorithm preventing long vehicle blockings and solutions to idle vehicles blocking others are required for bidirectional AGVSs. A symbiotic simulation resources scheduling decision support method is developed, including identification of route segments where deadlocks can potentially occur and an accordingly integrated banker's algorithm. Based on multiple online what-if simulations, the best deadlock safe resources schedule is determined for each short period just before the vehicles move. This proactive what-if analysis of resources utilization, alternative routes and dynamic parking strategies allows the minimization of the vehicles' total blocking time. Combining symbiotic simulation and a real time control of autonomous guided vehicles increases the AGVS's efficiency -- in terms of deadlock safety and minimizing the total vehicle blocking time. This method enables AGVS performance evaluation under arbitrary dispatching, routing, scheduling and guide-path design strategies.