Using Directed Acyclic Graph Support Vector Machines with Tabu Search for Classifying Faulty Product Types

摘要

Diagnosing quality faults is one of the most crucial issues in manufacturing processes. Many techniques have been presented to diagnose fault in manufacturing systems. The S VM approach has received more attention due to its classification ability. However, the development of support vector machines (SVM) in the diagnosis of manufacturing systems is rare. Therefore, this investigation attempts to apply the SVM in the diagnosis of manufacturing systems. Furthermore, the tabu search is employed to determine two parameters SVM model correctly and efficiently. A numerical example in the previous literature was used to demonstrate the diagnosis ability of the proposed DSVMT (directed acyclic graph support vector machines with tabu search) model. The experiment results show that the proposed approach can classify the faulty product types correctly.