摘要
针对河北某钢厂连铸连轧生产线的批量调度问题,选取宽度、厚度及硬度作为跳跃惩罚函数,以最小跳跃惩罚、最大轧制距离和最小空闲惩罚作为优化目标,建立热轧批量计划多目标模型。为避免轧制计划目标函数的数量级不同引起的加权系数选取困难,将热连轧批量计划问题转化为多目标优化问题,并采用自适应离散差分进化算法进行优化。引入了考虑偏好信息的最优解选取方法,为工厂生产提供多样化的解决方案。通过离散差分进化算法优化后的批量计划,减少了热轧批量计划中的宽度、硬度、厚度跳跃,有助于提高生产效率。
Focusing on the batch scheduling problem of continuous casting and rolling production line in a steel mill of Hebei province, the width, thickness and hardness were selected as the jump penalty function, and the multi-objective model of hot rolling batch schedu- ling was established by taking minimum jump penalty, maximum rolling length and minimum idle time as optimization objectives. To avoid the selection of weight coefficient caused by the different magnitude, hot rolling batch planning problem was transformed to muhi-objective optimization problem. The self-adaptive discrete differential evolution algorithm was employed to optimize the problem. The solution selec- tion method based on the preference was introduced to provide a variety of final solution for the factory. The batch planning based on dis- crete differential evolution algorithm can reduce the width, thickness and hardness jump and improves the production efficiency.
出处
《塑性工程学报》
CAS
CSCD
北大核心
2017年第3期148-153,共6页
Journal of Plasticity Engineering
基金
河北省自然科学基金资助项目(F2016203249)
河北省高等学校创新团队领军人才培育计划项目(LJRC013)
国家自然科学基金委员会与宝钢集团有限公司联合资助项目(U1260203)
河北省研究生创新资助项目(CXZZBS2017049)
关键词
热连轧
批量计划
多目标优化
差分进化
hot tandem rolling
batch scheduling
muhi-objective optimization
differential evolution
