Neural-expert hybrid approach for intelligent manufacturing: A survey

The development of computer-aided manufacturing systems is evolving towards that of intelligent manufacturing systems (IMS). In an intelligent manufacturing system, unprecedented and unforeseen situations are expected to be solved within certain limits, even on the basis of incomplete and imprecise information. Therefore, a tremendous amount of manufacturing knowledge is needed. Artificial intelligence (AI) based techniques are designed for capturing, representing, organizing, and utilizing knowledge by computers, and hence will play an important role in intelligent manufacturing. As an AI technique, knowledge-based expert systems have been used in manufacturing for nearly two decades. Recently, another AI technique, namely, neural networks, is gaining more and more visibility and has been successfully applied in manufacturing practice. Expert systems and neural networks are individually useful but not sufficient on their own in dealing with manufacturing problems. In fact, expert systems and neural networks are complementary. The weaknesses of expert systems are offset by the strengths of neural networks, and vice versa. Quite a few researchers are interested in merging the structures and functions of expert systems with those of neural networks. Some evidence has shown that the neural-expert hybrid technique is a promising tool for intelligent manufacturing. This paper reviews the current research on the neural-expert hybrid approach and its applications in manufacturing, which will provide some guidelines and references for the research and implementation.