摘要
针对人工蜂群算法收敛速度慢、容易出现"早熟"的缺点,提出了一种混合的人工蜂群算法(hybridartificial bee colony,HABC)。在人工蜂群算法的迭代中引入淘汰规则和新的搜索策略,以提高算法的收敛速度;同时,为了维护群体的多样性,对种群中的个体采用差分进化。通过对一个调频(frequency-modulated,FM)合成器参数优化问题测试,表明该算法能够有效地克服"早熟"现象,提高了全局寻优的能力。将其应用于线性系统逼近问题,仿真实验表明该算法是快速有效的。
In order to overcome prematurity and low searching speed of the artificial bee colony（ABC）,a hybrid artificial bee colony（HABC）algorithm is proposed.An eliminative rule and the new search strategy is introduced into the iteration of ABC to improve the convergence rate.Then,to maintain the population diversity, differential evolution（DE）simulates evolution and all individuals are taken into account in each generation.One experiment of parameter optimization of frequency-modulated（FM）synthesis indicates that the proposed algorithm can avoid prematurity effectively,and the algorithm possesses better ability in finding global optimum than compared algorithms.The proposed algorithm can be used to solve linear system approximation problems, and results show that the algorithm is fast and effective,and greatly outperforms other algorithms.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2011年第5期1167-1170,共4页
Systems Engineering and Electronics
基金
国家自然科学基金(60974082)
中央高校基本科研业务费专项资金(K50510700004)资助课题
关键词
人工蜂群
差分进化
线性系统逼近
参数优化
artificial bee colony
differential evolution
approximation of linear system
parameter optimization
