Continuous annealing production scheduling in iron & steel industry

This paper studies continuous annealing production scheduling for coils which arises in integrated steel industry. This problem is to sequence the candidate coils with the objective of minimizing the total switching penalties, so that productivity and product quality is improved. The problem is formulated as a variant traveling salesman problem where the objective function includes the switching penalties incurred by three consecutive coils besides the classical switching penalties incurred by two adjacent coils. To solve this problem, a tabu search based algorithm is proposed. In implementing tabu search, when solution traps in local optima, alternate exchange neighborhood is searched to kick it to a profitable region. Computational results on practical production data set demonstrate that the proposed algorithm for the model can effectively reduce the switching number for two adjacent coils and smooth the switching trend on changeover for three consecutive coils. The comparison between the proposed algorithm and dynamic programming shows the proposed algorithm can solve continuous annealing production scheduling problem for coils to optima for some small actual instances.