A multiple-attribute method for concurrently solving the pickup-dispatching problem and the load-selection problem of multiple-load AGVs

The pickup dispatching and the load selection are two control problems in multiple-load AGVs. Although they affect each other and are affected by various attributes, many researchers have solved them as separate problems and adopted single-attribute methods for them. In this paper, we propose a multiple-attribute method that can solve them simultaneously. The proposed method has four stages: preparation, clustering, evaluation and execution. At the preparation stage, we calculate the weights for three attributes (i.e., slack time, waiting time and distance) that are important to our problems based on the system's current status. These weights will be useful at the second and third stages. At the clustering stage, parts needing vehicle service are clustered into part groups based on their similarity in these three attributes. At the evaluation stage, part groups are evaluated by considering these three attributes. The part group with the greatest evaluation value will be served by the AGV. At the execution stage, a procedure is proposed to assist the AGV in picking up parts efficiently. Simulations were conducted to test the performance of the proposed method in throughput, flow time, and tardiness. The results show that the proposed method outperforms not only single-attribute methods, but also methods that solve pickup dispatching and load selection separately.