Research on Knowledge Representation in Choosing Mechanical Transmission Type

Abstract

To tackle the knowledge representation problem in developing intelligent conceptual design system, a BP network based approach is proposed, in which the related knowledge regarded various mechanical transmissions can be acquired and represented with the trained weight and threshold of neural network if they are turned into numerical data, encoded with binary strings, employed as learning samples, and fed into the constructed BP neural network. In a sense, the trained neural network can be used as a knowledge base of expert system to facilitate the choosing process of mechanical transmission. This paper provides a promising approach to deal with the automation of knowledge representation in conceptual design of mechanical transmission.