A Benchmark Study For Steelmaking-Continuous Casting Scheduling Optimization

Steelmaking - continuous casting scheduling optimization is a hot research topics, on which a lot of research was carried out by Tang Li-Xin , Li Tie-Ke and many overseas scholar such as Conway, Numao. Steelmaking - continuous casting scheduling is a NP-complete problem and the space of its solution is very big. Thus the computing speed of algorithms such as genetic algorithm and simulated annealing algorithm in the literature can not meet the requirements of schedule control.The three-constraint model of Steelmaking - continuous casting scheduling obtains three dimension decision space: the task assignment, the processing sequence and the time table. Combinatorial number of the first two dimensions is very huge when the processing route is complex, the size of continuous casting is large and the steel grade of each casting machine is different. To resolve this problem, dimension transform was proposeed in this paper, the dimension of the task assignment and the processing sequence are transformed into proper position in sequence to accept outsourcing and proper heat for outsourcing, which greatly reduced the feasible solution space.Schedule evaluating indicator was proposed in this thesis. The continuous casting scheduling is a job grouping problem. Therefore the continuity is the most important evaluating indicator. The converter is a bottle neck, and the converter load's proportionality is the second evaluating indicator. The equipment's rhythmicity and the workpiece's rhythmicity are the support conditions of schedule robustness. Three evaluation indicator was effectively used in choosing the optimal solution.This research was the basic research, which provides a benchmark for the schedule control system to debug and to improve algorithm in Baotou Steel Factory. This schedule control system has already run into operation in Baotou Steel Factory sheet-metal mill.