Production-inventory scheduling using Ant System metaheuristic

The present paper presents the algorithmic solution, based on an Ant System metaheuristic, of an industrial production-inventory problem in a steel continuous-casting plant. The model proposed is based on an objective function, the aim of which is to find the most profitable production schedule of the steel billets. Furthermore, the model takes into account the relevant parameters of the finite-capacity productive system (e.g. set-up and processing times, demand profile, warehouse capacity). Moreover, the make-to-order production environment of the company presents a significant manufacturing phase, which is represented by the billet cooling warehouse (similarly to the drying process in paper and textile production, or maturing in food production): this fact introduces a relevant constraint to production schedule. The study shows the basic criteria used for the problem modelling and the steps proposed for profit optimisation. The Ant System algorithm implemented is discussed and its relevance for the steel plant production management is shown.