摘要
模拟生物界成长发育过程,加入成长算子对遗传算法框架进行改进,形成新的算法框架-成长遗传算法(growth GA)．该算法能够克服简单遗传算法寻优速度较慢、局部搜索能力较弱的缺点．利用爬山法局部搜索能力强的特点,给出成长算子的一种具体实现,并证明加入成长算子不改变算法收敛性．与简单遗传算法和确定性拥挤遗传算法的对比函数优化实验证明:成长遗传算法有利于兼顾寻优速度和收敛精度．
By emulating the process of growth in nature and using growth operator, a growth genetic algorithm （GGA） is proposed to overcome the drawbacks of simple GA （SGA） such as slow optimization speed and weak local search ability. A practical realization of growth operator is proposed by making use of the strong local search ability of the hill climbing method. It has been demonstrated that adding the growth operator doesn＇t change the convergence property of SGA. The simulation result compared with SGA and deterministic crowding GA （DCGA） for function optimization verifies that the growth genetic algorithm facilitates the balance between optimization speed and convergence precision.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2006年第5期815-818,共4页
Control Theory & Applications
基金
中国科学院知识创新工程重大项目(KGCX-SW-15)
安徽省优秀青年科技基金资助项目(04042046).
关键词
成长遗传算法
成长算子
收敛性
函数优化
growth genetic algorithm
growth operator
convergence property
function optimization
