炼钢连铸生产调度问题建模及改进遗传算法求解  

An improved genetic algorithm and mathematical programming model for scheduling steel-making continuous casting production

作　　者：张启敏[1,2] 唐秋华[1] 郑鹏[1] 张利平[1] 王春锋[2] 

机构地区：[1]武汉科技大学机械自动化学院,武汉430081 [2]武汉钢铁股份有限公司,武汉430081

出　　处：《现代制造工程》2016年第11期50-55,共6页Modern Manufacturing Engineering

基　　金：国家自然科学基金资助项目(51275366;51305311;50875190);教育部博导和博士后科学基金项目(20134219110002;2013M542073)

摘　　要：炼钢连铸是钢铁制造的核心环节,多阶段、多约束及多缓冲的特性决定其生产调度的复杂性,研究高效稳定的生产调度方法有益于提升生产效率、减少能耗和降低成本。基于浇次间-浇次内两层分解,建立了基于单元特定事件连续时间的数学规划模型,并提出相应的算法进行求解。在浇次内,根据约束满足要求,提出与工艺时间相关的浇次内最大炉次数限制条件。在浇次序列求解中,根据问题特性设计了一种改进遗传算法,算法通过加入自适应遗传算子和基于邻域搜索的变异操作,以提高搜索性能。最后通过实际算例,验证了算法在复杂问题中的有效性和优越性。Steelmaking and Continuous Casting （SCC） is the core part of the steel production process. The characteristics of multi-constraint, multi-stage and multi-buffer complicated its production and scheduling. Therefore, researches on the production scheduling problem of SCC manufacturing process may improve the production efficiency, decrease the energy consumption and reduce the production costs. Based on the two layers of decomposition of in-cast and between-cast,its mathematical programming model is established based on unit-specific event-point continuous-time representation and the corresponding algorithm is proposed to solve it. In cast, the maximum number of ladle constraint which is related to process time is proposed according to the constraint satisfaction conditions. In order to obtain the cast sequence, an Improved Genetic Algorithm （IGA） is designed. In IGA, the self-adaptive genetic operators and neighbourhood-based mutation operation are adopted to improve its performance. Finally, through the experimental comparison,the effectiveness and efficiency of IGA in solving the large-size problems is demonstrated.

关 键 词：炼钢连铸 浇次 炉次 遗传算法 

分 类 号：F224[经济管理—国民经济]