| Integrated steel companies face an economic environment with strong competition, shrinking markets, and lagging profitability. Improved production and strategic planning are clearly of extreme importance. This dissertation develops several models that assist production and strategic planning in the steel industry.; A multi-period steel optimization model with both facility loading and costing functions was developed. In this model, the complicated steel production process was represented by a production network composed of facility nodes, material nodes, and material flow arcs. Based on the network representation, the steel production process was formulated as a linear programming problem with the objective of profit maximization. Its optimization and LP features make comparing different strategic and tactical alternatives possible by either running different scenarios or performing parametric analyses.; The importance of inventories, i.e., linkages between time periods in the steel optimization model, was also investigated. Data for plate mills and batch annealing processes of Bethlehem, ARMCO, and LTV steel companies were used in an empirical study. The empirical relationships between throughput and inventory level for plate mills and batch annealing processes were successfully determined by both least squares (LS) and least absolute deviation (LAD) regressions. A steel inventory model that performs LAD regressions to determine the empirical productivity versus inventory relationship was developed using 4th Dimension database software.; Furthermore, the issue of plate mill scheduling was also studied. The plate mill production scheduling work can be accomplished by two steps: first assign slabs to each plate order and then sequence the rolling jobs. Slab assignment was formulated as an LP problem with the objective of either maximizing yield, maximizing revenue, or maximizing profit. Both the optimal slab assignment and the optimal slab inventory mix are determined by the slab assignment model. The job sequencing problem is to find a job sequence that fulfills operational constraints and also maximizes plate quality. A network job sequencing model was developed. The model is able to efficiently (in polynomial time) sequence all the rolling jobs within a full roll change cycle. |
