Abstract: The reasonable classification of iron and steel scrap is the basic link of the delicacy management in the reverse supply chain for the iron and steel industry.For making full use of scrap iron and steel,the cyclic utilization is analyzed and a classification model based on the source,geometry appearance and chemical composition is proposed for the scrap iron and steel in the reverse supply chain.With the low versatility of the existing classification and the huge data of the chemical component,the dimension matrix composition data is reduced by the using of the principal component analysis at first.And then the initial cluster number and initial clustering result is obtained by hierarchical clustering method.Finally,a suitable classification has been constructed.In the end,the initial clustering result is optimized by applying ant colony algorithm.In a case of classification of 413 types of scrap iron and steel,9 kinds of chemical elements are regarded as an index of calculation.Finally,these scrap iron and steel can be classified in 5 species of scrap iron and 7 species of scrap steel.The chemical composition of each species is significantly different.This conclusion can solve the problems of warehouse layout and process decision and production scheduling in the iron and steel industry.

Key words: scrap iron and steel , reverse supply chain, classification research