论文部分内容阅读
针对模糊聚类算法不适应复杂环境的问题,提出了一种新的动态进化聚类算法,克服了传统模糊聚类建模算法须事先确定规则数的缺陷。通过改进的遗传策略来优化染色体长度,实现对聚类个数进行全局寻优;利用FCM算法加快聚类中心参数的收敛;并引入免疫系统的记忆功能和疫苗接种机理,使算法能快速稳定地收敛到最优解。利用这种高效的动态聚类算法辨识模糊模型,可同时得到合适的模糊规则数和准确的前提参数,将其应用于控制过程可获得高精度的非线性模糊模型。
Aiming at the problem that the fuzzy clustering algorithm does not adapt to the complex environment, a new dynamic evolution clustering algorithm is proposed, which overcomes the defect that the traditional fuzzy clustering algorithm needs to determine the number of rules in advance. The chromosome length is optimized through improved genetic strategy to achieve global optimization of the number of clusters. FCM algorithm is used to accelerate the convergence of clustering center parameters and the memory function and vaccination mechanism of the immune system are introduced to make the algorithm fast and stable Convergence to the optimal solution. By using this kind of efficient dynamic clustering algorithm to identify the fuzzy model, we can obtain the appropriate fuzzy rules number and accurate precondition parameters at the same time, and apply it to the control process to get the high-precision nonlinear fuzzy model.