摘要
为解决磷虾觅食(KH)优化算法在处理高维多模态函数优化问题时存在局部搜索能力不强、收敛速度慢等问题,利用一种贪婪的精英交叉算子加速其收敛速度,使用基于逻辑自映射函数的混沌搜索算子避免局部极值的吸引,采用对立搜索算子提高初始种群的质量。结合上述3种算子提出一种改进的磷虾觅食算法。在7个标准测试函数上的仿真实验结果表明,与KH及其改进算法相比,该算法在寻优精度和收敛速度方面均得到明显增强。
Krill Herd(KH)foraging optimization algorithm is one of the most recent achievements in the field of bionic swarm intelligence. Despite high performance of KH,weak local searching ability and slow convergence speed are two probable deficiencies for solving some high-dimension and multi-modal function optimization. This paper proposes a greedy elitism crossover operator for accelerating convergence,utilizes one chaotic searching operator to escape some local optima based on self-logical mapping function,and employs an opposition searching operator to improve quality of initial population. A new improved KH algorithm combining such three operators is given. Simulation results on7 benchmark functions show that the new algorithm has remarkable global optimizing ability and fast convergence speed,and outperforms the original KH algorithm and its newest variant algorithm.
出处
《计算机工程》
CAS
CSCD
北大核心
2015年第3期156-161,共6页
Computer Engineering
基金
中央高校基本科研业务费专项基金资助项目(JBK130503)
四川省教育厅基金资助项目(14ZB0046)
教育部人文社会科学研究基金资助项目(10YJCZH153)
关键词
磷虾觅食算法
局部搜索能力
对立策略
精英交叉算子
混沌搜索
收敛速度
Krill Herd(KH)foraging algorithm
local searching ability
opposition strategy
elitism crossover operator
chaotic searching
convergence speed