Using unsupervised learning technologies to induce scheduling knowledge for FMSs

This paper aims at applying unsupervised learning techniques to obtain scheduling knowledge for flexible manufacturing systems. The purpose of the research here is mainly to search for the inverse relationship between the operating performance and its corresponding decision parameters in the system. It may be explained that the learning effort tries to obtain the scheduling knowledge which determines the decision-mix to control the system to satisfy a certain required system performance. In the research, a typical flexible manufacturing system is configured to help illustrate the research procedures. In the modelled system, a set of decision parameters and system performance evaluation criteria were defined. The learning work starts with the collection of simulated data through the system, then groups the data into classes. Finally it applied the classification tree construction technology to obtain the definition of each class. The definitions are converted into rule-type knowledge, and thought was used as the scheduling knowledge.