基于混合遗传算法的热轧生产调度优化方法  

Development of Hybrid Genetic Algorithm for Multiple Round Scheduling in Hot Strip Mill

作　　者：邓宇巍[1] 吕勇哉[1] 陈玉旺[1] 

机构地区：[1]上海交通大学自动化系,上海200240

出　　处：《控制工程》2007年第B05期67-69,87,共4页Control Engineering of China

摘　　要：热轧生产调度是一个复杂的约束组合优化问题,其生产约束包括连续轧制板坯的宽度、厚度和硬度跳变要求,轧制单元的最大长度,产品库存及交货期等。基于多旅行商模型,建立了热轧生产批量调度问题的优化模型,并提出一种混合遗传算法(遗传算法、局部搜索)求解该问题。通过应用串行边重组和并行边重组的遗传交叉算子,算法在优化过程中可以很好地处理调度约束。针对工业数据的仿真结果证明该调度模型和混合遗传算法的并行求解策略可以有效地解决热轧生产批量调度问题。The development of building a muhi-round scheduling （MRS）solution for hot strip mill （HSM）with the hybrid genetic algorithms （HGA）is presented. The MRS is formulated as a constrained optimization problem with its desired criterion subject to variety of constraints, such as the patterns of width, gauge and hardness, groove, min-max rolling capacity, inventory and due delivery date, etc. The HGA is developed in terms of the combination of GA with a local search algorithm （GA-LS）to solve this complicated scheduling problem. The HGA employs a set of special genetic operators, such as SERX （serial edge recombination crossover）and PERX（parallel edge recombination crossover）to deal with the constrained evolutions and to produce several rounds with parallel strategy. The simulation results show the proposed GA-LS can solve multiple round parallel scheduling problem effectively.

关 键 词：热轧生产调度 遗传算法 局部搜索 边重组 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]