基于基因库和DNA重组技术的带钢层流冷却系统多目标优化  

Strip steel laminar cooling system′s multi-objective optimization based on gene pool and DNA recombination technology

作　　者：孙铁军[1] 曲丽萍 刘冲杰 路赵 SUN Tiejun;QU Liping;LIU Chongjie;LU Zhao(College of Electrical and Information Engineering,Beihua University,Jilin 132021,China;Engineering Training Center,Beihua University,Jilin 132021,China)

机构地区：[1]北华大学电气与信息工程学院,吉林吉林132021 [2]北华大学工程训练中心,吉林吉林132021

出　　处：《现代电子技术》2020年第11期101-105,共5页Modern Electronics Technique

摘　　要：针对层流冷却系统粗调区目标卷取温度和给定冷却速率的二维多目标优化问题,提出基于基因库和DAN重组技术优化的多目标遗传算法来获取粗调区集管的最佳开闭模式群。该算法利用Pareto前沿面的交集建立基因库,从中挖掘出集管开闭的较优模式,将其耦合至下一代种群,最大限度地消除了种群进化的随机性和漫游性。基因库的优胜劣汰机制,有利于保持种群个体在搜索空间的分布多样性,使算法可以在更广阔的空间搜索出更佳的集管开闭模式。基因库的随机抓取策略保证了Pareto前沿面在全局搜索空间的均匀分布性,增强了控制系统对多目标的均衡控制能力。最后,基于DNA重组技术强力驱动算法收敛于基于全局的最优目标解群,极大限度地提高了控制系统的控制精度。编写了基于微软基类库的仿真程序,仿真结果验证了该多目标优化策略的先进性。In allusion to the two-dimensional multi-objective optimization of the objective coiling temperature and the given cooling rate in the coarse adjustment area of laminar cooling system,a multi-objective genetic algorithm optimized with the gene pool and the DAN recombination technology is proposed to obtain the optimal opening and closing mode group for the manifold in the coarse adjustment area. In the algorithm,the intersection of Pareto frontier is used to build the gene pool,from which a better opening and closing mode of the manifold is mined and then coupled to the population of next generation to eliminate the randomness and roaming of population evolution to the greatest extent. The mechanism for the survival of the fittest of the gene pool is beneficial to maintaining the diversity of individual population distribution in the search space,so that a better opening and closing mode of the manifold can be searched out with the algorithm in a broader space. The random grabbing strategy of the gene pool ensures the uniform distribution of Pareto frontier in the global search space,and enhances the control system′balancing control to multiple objectives. The DNA recombination technology based algorithm with strong driving mechanism converges to the global optimal objective solution group,which greatly improves the control accuracy of the control system. The simulation program based on the Microsoft foundation class(MFC) library was compiled. The simulation results verifies the advancement of the multi-objective optimization strategy.

关 键 词：多目标优化 带钢层流冷却 Parteo前沿面 基因库 DNA重组技术 均衡控制 

分 类 号：TN911.1-34[电子电信—通信与信息系统] TP273[电子电信—信息与通信工程]