用P-MA解决多目标炉次计划编制问题

来源期刊：冶金自动化2009年第2期

论文作者：张健欣 童朝南

文章页码：7 - 11

关键词：炉次计划编制； 多目标优化； Memetic算法； 邻域搜索算法； 进化算法

Key words：heat scheduling； multi-objective optimization； Memetic algorithm； neighborhood search algorithm； evolutionary algorithm

摘    要：针对炼钢-连铸-热轧一体化生产中的炉次计划编制问题建立了基于多目标优化的数学模型,并设计了P-MA(Pareto-based Memetic Algorithm)。基于多目标优化的数学模型能更好地反映炉次计划的工艺约束,可以得到多个Pareto非支配解,为一体化生产的动态协调和调整提供了更多的炉次方案,使整个生产过程具有更大的灵活性。P-MA采用SPEA2的框架并结合了基于Pareto支配性的自适应网格存档(AGA)策略来评价邻域解的质量,以确定是否接受这个新解。通过对实际生产数据的仿真实验验证了P-MA在求解多目标炉次计划编制问题上是有效的。

Abstract: For heat scheduling problem of steelmaking-continuous casting-hot rolling integrated production,a mathematic model based on multi-objective optimization(MOO) was built,and P-MA(Pareto-based Memetic Algorithm) was designed.The mathematic model based on multi-objective optimization can reflect process constraints well,and many Pareto non-domination solution can be obtained,so that many heat plans can be presented for dynamic coordination and adjustment of integrated production,and production process was more flexible.P-MA with SPEA 2 frame was combined with AGA(Adaptive Grid Archiving) strategy to evaluate quality of neighborhood solution.Practical data from a steel plant is cited for demostrating effectness and efficiency of the model and algorithm.The results show that the algorithm can achieve satisfactory efficiency.