Production-mix optimization of a syderurgic plant in relation to market and plant set-up

The management of industrial plants involves difficult decision problems. First of all it is necessary to find a plant set-up giving a good return on investment and a competitive cost of production. Then it is necessary to find ‘what and how much’ to produce in order to define an optimum production mix, according to market demand and capacity. Moreover, it is often necessary to face a complex structure involving more production plants: the target is to get an efficient management of the whole system. This must be obtained by splitting production among the different plants and by optimizing exchanges of raw materials and semi-finished products among the production centres-that gives a minimum transport cost. Our study faces this kind of problem. In particular, a DSS for the short and medium term management of an industrial plant has been defined. This method of system analysis uses mathematical optimization tools (such as linear programming) that have been integrated by parametral analysis, multicriteria analysis and cluster analysis. The problem was very complex, because it had some thousands of constraints and variables and a lot of hypotheses to be analyzed. This fact made necessary a main frame computer and a high level software.