摘要
杂草算法是受杂草扩张繁殖行为启发而来的一种新颖的仿生群智能优化算法。在分析基本杂草算法仿生原理和局限性的基础上,提出一种改进的入侵杂草优化算法,运用混沌反向学习策略对种群进行初始化,通过变异算子增加种群的多样性;并对种群中的精英个体进行混沌搜索,以提高其跳离局部最优值的能力。对经典函数的仿真测试表明,改进算法性能优于基本杂草算法,是解决工程应用复杂函数优化问题的一种有效方法。
Inspired by the reproductive aggressive behavior of weeds in nature, invasive weed optimization algorithm (IWO) was developed as a novel bionic swarm intelligence optimization algorithm. An improved IWO algorithm was proposed on the basis of analyzing bionic principle and limitations of basic IWO, which applied an initialization strategy based on chaotic opposition-based learning, increased the diversity of the population through the mutation operator, and enhanced its ability to jump out of local optimal value by chaotic search around current elites. Simulation results for benchmark functions show that the proposed algorithm has improved optimization property compared with IWO, as an effective method to solve complex function optimization problems in engineering application.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2016年第8期1732-1739,1747,共9页
Journal of System Simulation
基金
国家自然科学基金(71271138)
上海市一流学科建设(S1201YLXK)
沪江基金(A14006)
关键词
杂草算法
仿生原理
混沌变异
仿真测试
invasive weed algorithm
bionic principle
chaotic mutation
simulation test