(A) study on operation problems for automated guided vehicle systems using fuzzy set theory

Abstract

Automated Guided Vehicle Systems(AGVS) have been widely used in manufacturing systems for material handling. This thesis is concerned with the dispatching and routing problem of the vehicles which has been recognized as a key element for the efficient operation of AGVS. In the first part of this thesis, a new AGV dispatching algorithm is developed which is flexible and adaptable to the system configurations. The algorithm, utilizing the fuzzy multi-criteria decision-making method, prioritizes all move requests based on the urgency of each move request. The following four kinds of dispatching criteria are considered simultaneously in measuring the urgency: elapsed time of move request, incoming buffer queue length, outgoing buffer queue size and AGV traveling distance. The performance of the proposed rule is compared with several well known dispatching rules in a hypothetical job shop environment in terms of the system throughput and makespan. Simulation experiments are carried out varying the capacity of buffer queues and the number of AGVs. The second part presents an adaptive dispatching algorithm by adding an adaptive control scheme to the proposed algorithm. In the adaptive scheme, the urgency of each move request is determined such that it can respond to the dynamic conditions of AGVS operation. At each dispatching decision point, the importance weights for dispatching criteria are adjusted by the adaptive scheme utilizing the informations related to the degree of urgency for the vehicles and the degree of congestion of incoming buffer. A comparative study is carried out by simulation experiments varying the following factors; queue capacity, dispatching criteria, the level of criticality ratio of AGVs, number of AGVs, and maximum level of work-in-process. The final part deals with the AGV which can carry more than one unit load. Firstly, a dispatching algorithm is developed for the system which operates only one vehicle. The algorithm prioritizes all move...