Data-mining approach to production control in the computer-integrated testing cell

This paper presents a data-mining-based production control approach for the testing and rework cell in a dynamic computer-integrated manufacturing system. The proposed competitive decision selector (CDS) observes the status of the system and jobs at every decision point, and makes its decision on job preemption and dispatching rules in real time. The CDS equipped with two algorithms combines two different knowledge sources, the long-run performance and the short-term performance of each rule on the various status of the system. The short-term performance information is mined by a data-mining approach from large-scale training data generated by simulation with data partition. A decision tree-based module generates classification rules on each partitioned data that are suitable for interpretation and verification by users and stores the rules in the CDS knowledge bases. Experimental results show that the CDS dynamic control is better than other common control rules with respect to the number of tardy jobs.