Abstract: To solve the difficulty of low reuse value in the traditional process route discovery methods based clustering analysis and to make the extracted process routes support the effective reuse based manufacturing resources, a multi-dimensional manufacturing information based typical product process route discovery method is presented. A process information element and process route information model based multi-dimensional manufacturing information are established, and consequently the lower-dimensional process route model of its own dimensionality is obtained by using kernel principal component analysis(KPCA) to reduce the dimensionality of process information element. Based on the lower-dimensional process route model, a distance calculation method for calculate the similarity between process routes is proposed and evolutionary cellular learning automata was applied to realize the intelligent clustering division of process routes. The typical process routes are extracted from the clustering clusters consequently. Experimental results show that the effectiveness of proposed method is verified.

Key words: clustering analysis, dimensionality reduction, evolutionary cellular learning automata, multi-dimensional manufacturing information, typical process route