关键词: 钢铁生产, 优化模型, 动态调度, 案例推理, 模型转换, 遗传算法

Abstract: With the purpose to optimize dynamic scheduling in steel-making, a CBR (case-based reasoning) system with embedded adaptation model or CBR/EAM is established, of which the function and establishing process of the model, especially how to solve it are presented. To solve it easily, the nonlinear adaptation model is converted into linear one. Then a two-stage algorithm is utilized to improve solving efficiency. First, the genetic algorithm (GA) that introduces heuristics to generate population is adopted to pretreat the problem so as to reduce relevant dimensions, the 3-dimensional natural number encoding is adopted to meet the requirement of practice. Then, the simplified model is solved with simplex method. The simulation result is verified that the validity of CBR/EAM is better than that of single model approach or conventional CBR approach to solve the dynamic scheduling problem in steel-making process. With the prototype system further perfected, it could be put into application with benefits expected.