Abstract
Modern manufacturing systems have incorporated various “intelligent” decision support techniques and algorithms in the form of rules-based systems which have been promoted extensively because of the recent development in fuzzy inference, artificial neural network, machine learning and optimization algorithms, and etc. The basic elements of a modern manufacturing system include design automation (CAD and CAPP), production automation (CAM and MRP II), and shop floor automation (SFC and FMS), which must then be integrated into a coherent system through information automation (distributed DBMS and workflow automation) to form a fully automated factory. Therefore, considerable amount of rule-based systems have been developed and implemented fully or partially into each of the above areas of automation. To build a robust rule-based system requires a suitable modeling and analysis tool and a systematic design methodology. By incorporating neural network concept and associated learning algorithm, we present in this paper a Petri net based modeling methodology which can be used to design and implement adaptive rule-based systems for control and optimization in intelligent manufacturing systems. Demonstration is done on parametric design and optimization in manufacturing processes.