| There are four commonly employed strategies for prepositioning idle vehicles in an AGVS. These strategies involve positioning vehicles at unload points, a central dwell point, circulatory loops, and at distributed dwell points. To what extent does the management of idle vehicles and dwell point specification affect a system's performance? Are some prepositioning strategies superior to others? The need to provide some answers to these questions is the focus of this research.;No computational effort is required to determine dwell points if idle vehicles are positioned at unload points. Also, a central dwell point is a special case of distributed dwell points. Therefore, to investigate the influence of prepositioning strategies on the performance of a shop, idle vehicle distributed dwell zone selection problem and circulatory loop selection problem were first formulated and solved. Based on the results of the optimization, AGVSim was used to simulate the operation of facilities based on various idle vehicle prepositioning alternatives.;Based on the simulation results, distributed dwell zones performed the best in most cases. Moreover, the analysis also revealed that idle vehicle prepositioning strategy can influence shop performance regardless of whether the shop is busy or not. However, prepositioning strategy does not affect, at least in the tested case, the performance of a shop in a bidirectional AGVS.;Several benefits can be derived from this study. First, the models developed could help AGVS designers to determine idle vehicle dwell zones or circulatory loops before AGVS installation. Second, more reliable performance prediction of an AGVS could be accomplished because empty vehicle flow is considered in the design of the system. Third, the models developed are useful for managing idle vehicles in order to improve job throughput time.;The contributions of this research can be summarized as follows. First, procedures were developed to characterize AGVS networks to determine potential idle vehicle dwell points. Second, heuristics for selecting dwell points were developed. Third, techniques for generating all loops were developed and reduction rules were proposed to ease computational effort for selecting loops. Fourth, the effectiveness of various idle vehicle prepositioning strategies were characterized. |
