烧结钕铁硼成型生产计划编制的混合优化算法  

Hybrid optimization algorithm for production planning of sintered NdFeB molding

作　　者：刘业峰[1] 柴天佑[2,3] 

机构地区：[1]沈阳工学院信息与控制学院,辽宁抚顺113122 [2]东北大学自动化研究中心,辽宁沈阳110819 [3]东北大学流程工业综合自动化国家重点实验室,辽宁沈阳110819

出　　处：《南京理工大学学报》2016年第6期679-686,共8页Journal of Nanjing University of Science and Technology

基　　金：国家自然科学基金(61603262;61403071);沈阳工学院博士启动基金(BS201503)

摘　　要：为提高烧结钕铁硼成型工序生产效率,建立了1个多目标混合整数规划模型。提出遗传算法和对偶单纯形算法相结合的混合求解算法,改进了遗传算法的交叉算子。采用3PM交叉算子和随机动态位串变异的方式,改进了遗传算法的变异算子。基于所建立的多目标优化模型和提出的求解算法编制了成型生产计划。仿真结果表明,较采用人工编制的生产计划,采用该文算法编制的成型生产计划,总产量增加了近500模,成型机利用率提高了15%,且每日进炉的炉重得到了增加。A multi-objective hybrid integer programming model is proposed to improve the efficiency of NdFeB molding process production. The crossover operator of a genetic algorithm is improved using a hybrid algorithm combining a genetic algorithm and dual simplex algorithm. The mutation operator of the genetic algorithm is improved using 3 parents method （ 3 PM ） cross operator and the dynamic variation of the random bit string. A forming production plan is established based on this multi-objective optimization model and the solution method. The simulation results show that, compared with the man-made production plan, the total output of the production plan of this algorithm increases about 500 die, the utilization rate of molding machines improves by 15% , and daily furnace weight increases too.

关 键 词：烧结钕铁硼 生产计划 遗传算法 随机动态位串变异 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]