摘要
提出一种基于云模型的入侵杂草优化算法,根据杂草适应度值的大小将杂草种群分为优良子群、普通子群和较差子群。通过CR调整标准差,不同的子群采取不同的标准差进行扩散,优良子群采用较小的标准差进行精细搜索,普通子群利用云模型的随机性和模糊性动态调整标准差,进行自适应搜索,较差子群采用较大的标准差进行全局搜索。由此加快了算法的收敛速度,较好地平衡了全局搜索能力和局部搜索能力,并且在一定程度上避免了算法陷入局部最优。对7个测试函数进行仿真实验,结果表明,该算法具有较高的寻优精度和更快的收敛速度。
In this paper,a kind of Invasive Weed Optimization（ IWO） algorithm based on cloud model is proposed. The weeds are individed into excellent group,the normal group and the poor group. The algorithm adjusts standard deviation through CR and different subgroup adopts different standard deviations to diffuse. The excellent group adopts the smaller standard deviation,which realizes the fine-grained search. The normal group applies the fuzziness and randomness of the cloud model to dynamically adjust the standard deviation, which realizes adaptive search. And it takes larger standard deviation for the poor group to do global search. This approach improves convergence rates,balances abilities of global search and local search better,and avoids the algorithm trapping in local optimum to some extent. Experimental results of several test function indicate the new algorithm has a better optimal performance.
出处
《计算机工程》
CAS
CSCD
2014年第12期156-160,共5页
Computer Engineering
基金
国家自然科学基金资助项目(61063028)
甘肃省教育信息化发展战略研究基金资助项目(2011-2)
关键词
杂草优化算法
云模型
精细搜索
自适应
局部最优
X条件云发生器
Invasive Weed Optimization（ IWO） algorithm
cloud model
fine-grained search
adaptive
local optimum
X-conditional cloud generator
