基于反向学习的差分进化算法的冷轧负荷分配  

A Differential Evolution Algorithm Based on Opposite Learning in Load Distribution for Cold Rolling

作　　者：赵志伟[1,2] 

机构地区：[1]唐山学院计算机科学与技术系,河北唐山063000 [2]燕山大学电气工程学院自动化系,河北秦皇岛066004

出　　处：《计量学报》2017年第4期453-458,共6页Acta Metrologica Sinica

基　　金：河北省高等学校科学技术研究青年基金项目(QN2017416);河北省科技计划项目(16212302);唐山学院博士创新基金

摘　　要：提出了一种基于反向学习的自适应差分进化算法,该算法为增加初始种群的多样性,采用反向学习技术进行种群初始化,然后根据选择概率为每一个个体选择变异策略,并使用基于进化的单调递减函数和Logistic映射为每一个个体产生控制参数值。经过大量的仿真实验,实验结果表明该算法要优于其他差分进化算法。最后将该算法应用到了冷连轧负荷分配优化中。A differential evolution algorithm based on opposite learning is presented. In the proposed algorithm, to increase the diversity of initial population, the opposite learning is employed. In addition, the mutation strategy assigned to each individual is adaptively selected according to the selected probability, and the control parameters are generated by evolution-based monotone decreasing function and Logistic mapping. A large amount of simulation experiments have been made. Experimental results show that the proposed algorithm is better than other differential evolution algorithms. At last, the proposed algorithm is applied in load distribution for tandem cold rolling.

关 键 词：计量学 差分进化 自适应 反向学习 负荷分配 冷连轧 

分 类 号：TB93[一般工业技术—计量学]