摘要
针对实际应用中大量存在的离散变量优化设计问题,研究了利用一般连续变量方法进行离散变量优化设计的不足。结合离散变量优化问题与遗传算法的特点,提出离散交叉算子和离散变异算子,使遗传算子真正在离散空间中进行搜索。基于线性搜索思想提出离散引导算子以提高遗传算法的局部寻优能力,将种群逐步向离散极值点进行引导,实现算法的快速离散寻优。通过对两个实际离散变量优化设计问题的应用研究,验证了本方法解决离散变量优化设计问题的有效性。
According to lots of discrete variable optimization problems in practice, the defects of applying continuous variable optimization to solve discrete variable optimization problems were studied. The characteristics of discrete variable optimization and genetic algorithm were associated. Thus, the discrete crossover operator and discrete mutation operator were proposed to make the genetic operator search in discrete space. Based on the theory of linear search, the discrete leading operator was proposed so as to improve the local searching capability of genetic algorithm, that led the population to local optimization and implemented rapid discrete searching. The study on two practical discrete variable optimization problems proves the validity of this algorithm in solving discrete variable optimization problems,
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2006年第5期1154-1156,共3页
Journal of System Simulation
基金
国家自然科学基金资助项目(60374003)
973子课题资助项目(2002CB312200)
教育部及辽宁省流程工业综合自动化重点实验室开放课题基金资助项目(PAL200509)
关键词
离散变量
遗传算法
离散交叉算子
离散变异算子
引导算子
discrete variable
genetic algorithm
discrete crossover operator
discrete mutation operator
leading operator
