Steel-making And Continuous Casting Planning Based On The Multi-objective Optimization

Iron and steel industry provides raw materials for other industries and plays anessential role in the industrial economy. Steelmaking and continuous casting is the keyprocedure of the steel-making flow. Therefore, making researches on planning andscheduling for steelmaking and continuous casting is of great importance in makingtechnical procedure clear, quickening technical rhythm, improving the enterprise’seconomic efficiency.This paper’s intent is to research the planning of steelmaking and continuous castingbatch production. The technology of multi-objective evolutionary algorithms is alsostudied. The main works in this paper are summarized as follows:A multi-objective optimization model of charge plan was built in this paper. Asdirect encoding is hard to be used in the many constraints problem case, a new method ofgenerate feasible solutions based on constraint satisfaction was proposed. A novel adaptivegrid algorithm was also proposed to improve the current multi-objective evolutionaryalgorithm’s population maintain strategy. Moreover, the proposed multi-objective chargeplan model was solved, and more meaningful solutions for this problem with subsequentscheduling were obtained. Finally, A programming model for the CAST plan is established.The computation with practical data shows that the model and the solving methodproposed are effective.In the study, a new composite evolutionary operator CSDO (Composite SA-SBXand Differential evolutionary Operator) for multi-objective evolutionary algorithm isproposed. It combines the self-adaptive simulated binary crossover operator anddifferential evolutionary operator, it guarantees the diversity of offspring and acceleratesthe exploration speed. Meanwhile, a “greedy” variable number offspring archive isproposed. It can improve the use efficiency of the evolutionary operators. The proposedmethods are combined to test the benchmark problems, and the results indicate that theaccuracy and searching speed are improved by using the proposed algorithm.