Developing Control Rules for an AGVs (Automated Guided Vehicles) using Markov Decision Processes.

An Automated Guided Vehicle (AGV) system is modeled using Markov Decision Processes. Generalized control rules for scheduling AGV's are extracted from the (Markov) optimal control policies. The job throughput (deliveries per time) using the extracted control rules is compared with that of the optimal policy and with those of control rules currently used in industry. Simulation results are provided to verify improvements achieved in other performance measures using the extracted rule for complex AGV systems.