Scheduling a Galvanizing Line by Ant Colony Optimization

In this paper, we describe the successful use of ACO to schedule a real galvanizing line in a steel making company, and the challenge of putting the algorithm to use in an industrial environment. The sequencing involves several calculations in parallel to figure out the best sequence considering the evolution of each important parameter: width, thickness, thermal cycle, weldability, etc. For solving this combinatorial (NP-hard) problem, new necessity arose to develop an intelligent algorithm able to optimize the scheduling, avoiding traditional manual calculations. Hence, ACO is proposed to translate the scheduling rules and current criteria into a set of technical constraints and cost functions to assure a good solution in a short calculation time.