改进型复杂过程全局进化算法在热连轧负荷分配中的应用

Load distribution optimization of hot strip mills with improved evolutionary algorithm for complex-process optimization

为实现热连轧精轧机组负荷分配的优化设定,提出一种具有柔性框架结构的改进型复杂过程全局进化算法.该算法部分地借用了分散搜索原则,在通用框架中嵌入具有搜索机制的子方法;利用无限折叠映射混沌模型和局部搜索法,分别对初始种群的生成和"超越"深度搜索进行改进以提高最优解的求解效率.实验结果表明,该算法能够使用较少的参数完成负荷分配优化问题的可行解搜索,具有较好的时效性,是局部和全局搜索的有机体.

An improved evolutionary algorithm for complex-process optimization（IEACOP） is presented to achieve optimal setting of load distribution about hot rolling mill. The algorithm is partially based on the principles of the scatter search, which has flexible structure and is embedded in sub-methods that have search mechanism. Meanwhile, infinite folding chaotic model and local search method are applied to improve initial population strategy and ＂go-beyond strategy＂ of in- depth search. Then the efficiency of the local optimal solution is improved. The experiment results show that IEACOP makes use of fewer adjustable parameters to get feasible mathematical solution for the actual load distribution problems and validate the real-time application, which is the organism including local search and global search.