摘要
针对磨矿过程的工艺特点,结合选矿厂对磨矿产品产量、质量的生产要求,提出磨矿过程多目标优化问题.为求解此多目标优化问题,研究一种快速的多目标遗传算法NSGA-Ⅱ(Non-dom inated Sorting Genetic A lgorithmⅡ).针对NSGA-Ⅱ算法中个体进行交叉前选择方法的不足之处加以改进,引入过滤、限制机制.仿真结果表明:引入过滤、限制机制可以限制"近亲"交叉,保持种群的均匀分布和多样性,加快种群在进化搜索过程中找到优秀个体的速度.将改进后的算法用于磨矿过程稳态优化,求得适合实际生产的多组系统操作参数,并用TOPSIS方法选出最优的一组操作参数.
Considering the technologic traits of grinding processes, combining the concentrator mill production＇ s output and quality, the multi-objective optimization problem of any grinding process is raised. To solve the problem of multi-objective optimization, a fast and outstanding multi-objective Genetic Algorithms NSGA-Ⅱ （Non-dominated Sorting Genetic Algorithm Ⅱ） was studied. The NSGA-Ⅱ was improved in a selection of individuals before crossover. The filtering and restriction mechanism were introduced and simulation results show that the introduction of filtering and restriction mechanisms may limit the crossover of ＂near relative＂, maintaining the uniform distribution and diversity of the population. It may also accelerate the process to find the best sample. The improved NSGA-Ⅱ was used for steady-state optimization of the grinding process. Many groups of suitable operating parameters of the actual production system can be found, and then the TOPSIS method can be used to select the optimal group of operating parameters.
出处
《沈阳化工学院学报》
2009年第1期55-60,共6页
Journal of Shenyang Institute of Chemical Technolgy
基金
辽宁省教育厅项目(2008566)
关键词
磨矿过程
多目标优化
遗传算法
非劣分层
grinding process
multi-objective optimization
genetic algorithm
non-dominated sorting
