带局部搜索的NSGAⅡ算法在负荷分配中的应用  

NSGAⅡ Algorithm Combined with Local Search and Its Application to Load Distribution of Hot Strip Mill

作　　者：李维刚[1,2] 贾树晋[2] 郭朝晖[2] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110819 [2]宝钢集团中央研究院自动化所,上海201900

出　　处：《控制工程》2012年第6期1011-1014,1018,共5页Control Engineering of China

基　　金：国家自然科学基金项目(50974039)

摘　　要：为了改善NSGAⅡ算法的收敛性能,将局部搜索的思想融入到NSGAⅡ中,提出一种带局部搜索的NSGAⅡ算法(NSGAⅡ-LS)。该算法采用基于惩罚的边界插入法(PBI)产生局部搜索的优化函数,并使用序列二次规划方法 (SQP)进行求解。通过在3个多目标标准测试函数上的仿真实验,结果表明相对于NSGAⅡ,NSGAⅡ-LS具有更好的收敛性能。最后将NSGAⅡ-LS应用到带钢热连轧负荷分配优化计算中,给出了兼顾轧制力平衡、最低轧制功率和优良板形的目标函数表达式,对多目标进化算法在热轧负荷分配优化计算中的应用进行研究,指出了目标之间的冲突关系。In order to improve the convergence performance of NSGA Ⅱ, a new multi - objective evolutionary algorithm, called NSGA Ⅱ , combined with local search （ NSGAⅡ - LS）, is proposed. In NSGA Ⅱ - LS, the penalty - based boundary intersection （PBI） ap- proach is used to give an optimization function for local search, and the sequential quadratic programming （SQP） algorithm is chosen as a local search solver. Simulations on several multi - objective benchmark instances indicate that the convergence performance of NSGA Ⅱ- LS is better than that of NSGA lI. Finally, NSGA Ⅱ - KS is applied to the load distribution optimization of a hot strip mill, with given expressions of objective functions including rolling force balance, rolling power minimizing and good strip shape. The multi - ob- jective optimization for draft scheduling has been performed, and the conflicting relation of 3 objective functions is analyzed.

关 键 词：多目标优化 局部搜索 NSGA-Ⅱ 负荷分配 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]