Article ID: | iaor20082665 |
Country: | Netherlands |
Volume: | 110 |
Issue: | 1/2 |
Start Page Number: | 213 |
End Page Number: | 223 |
Publication Date: | Jan 2007 |
Journal: | International Journal of Production Economics |
Authors: | Everett J.E. |
Keywords: | artificial intelligence: decision support, production |
Iron ore is railed to port, crushed to lump and fines, stockpiled and shipped. Product quality requires consistent composition, in iron and other minerals. Decision support systems described here aid daily ore selection, to maintain target composition. The method evolved from a batch system, building each stockpile to target, to a flow process, targeting an exponentially smoothed continuous stockpile. This has decoupled mine, rail and port operations, lessening costly rehandling. A quadratic programming solution guides ore selection. Some ill-defined and cost related constraints require human judgment. The emphasis is on decision-support, not decision-making, in a fruitful example of human/computer interaction.