摘要
针对模糊分类问题,提出了一种基于烟花算法和差分进化算法的模糊系统建模方法,这是首次将烟花算法应用到模糊建模领域,目的是建立精确度高而且解释性好的模糊分类器.烟花算法是通过模拟烟花爆炸现象而提出的一种新的粒子智能算法,首先采用此方法对模型的结构和参数进行学习,结果表明该方法具有较好的收敛速度和搜索性能;其次,为了扩大搜索范围,避免过早地陷入局部最优,在每一次迭代中采用差分算法进一步优化模型;最后,对Iris数据样本进行仿真实验,在保证较高分类精度的前提下,构建了一个解释性良好的模糊分类系统.
A novel approach to construct accurate and interpretable fuzzy classification system based on fireworks optimization algorithm（ FOA） combined with differential evolution operators is proposed. It is the first time to apply FOA in fuzzy modeling. The fireworks optimization algorithm is a novel swarm intelligent algorithm based on simulating the explosion process of fireworks,which can optimize the construction and parameters of fuzzy system with good convergence speed and optimization accuracy. To improve the diversity of the swarm and avoid being trapped in local optima too early,the differential evolution is performed to further optimize the model at each iteration. The proposed approach is applied to the Iris benchmark classification problem,and the results prove its validity.
出处
《郑州大学学报（工学版）》
CAS
北大核心
2015年第6期47-51,共5页
Journal of Zhengzhou University（Engineering Science）
基金
国家自然科学基金资助项目(61473266)
关键词
差分进化
烟花算法
模糊系统
解释性
系统辨识
differential evolution
fireworks algorithm
fuzzy system
interpretability
system identification
