A Decomposition Algorithm for Hot Rolling Planning

A fuzzy C-means clustering and constraint programming based multi-objective optimization decomposition algorithm for the hot rolling planning problem is presented. The slabs of a turn are divided into several clusters by fuzzy C-means clustering; sequence of slabs within each cluster and sequence of clusters are determined by constraint programming; the overall plan is obtained by integrating the sequence of slabs in accordance with the sequence of clusters. Computational experiments based on production data and random data show that the proposed algorithm performs well both in efficiency and effectiveness.