融合粗糙集和二元萤火虫算法的雾霾关键影响因素预测方法

发布时间：2018-01-09 18:00

本文关键词：融合粗糙集和二元萤火虫算法的雾霾关键影响因素预测方法　出处：《系统工程理论与实践》2017年01期 　论文类型：期刊论文

【摘要】：雾霾对人类的日常生活带来极大的危害,因而分析产生雾霾的关键影响因素尤为重要.针对目前传统算法预测雾霾关键影响因素存在的缺陷,从一维细胞自动机入手,提出了一种以基于群落弱连接机制的二元萤火虫算法(CWLBGSO)为搜索策略,粗糙集为评价准则的混合方法.CWLBGSO基于自然界中萤火虫间协同进化的"弱连接"机制,划分搜索空间,为每个子空间分配相应的种群,各子种群中的次优个体相互交互产生新个体,从而保持种群的动态多样性,然后将CWLBGSO结合粗糙集,应用于北京,广州和上海三地雾霾关键影响因素的预测中,并结合10交叉验证和SVM算法对预测结果分类准确率和影响因素进行分析,通过与其它算法进行对比,结果表明本文算法能有效剔除冗余因素,预测结果具有较高的稳定性和可行性.
[Abstract]:Haze brings great harm to human daily life, so it is very important to analyze the key influencing factors of haze. Based on one-dimensional cellular automata, a two-variable firefly algorithm, CWLBGSO, based on community weak linkage mechanism, is proposed as a search strategy. CWLBGSO. Based on the "weak connection" mechanism of coevolution between fireflies in nature, CWLBGSO divides search space and assigns corresponding populations to each subspace. The sub-optimal individuals in each sub-population interact with each other to produce new individuals so as to maintain the dynamic diversity of the population. Then CWLBGSO combined with rough sets is applied to Beijing. In the prediction of key influencing factors of haze in Guangzhou and Shanghai, combined with 10 cross-validation and SVM algorithm, the classification accuracy and influencing factors of prediction results are analyzed, and compared with other algorithms. The results show that the proposed algorithm can eliminate redundant factors effectively and the prediction results have high stability and feasibility.
【作者单位】：合肥工业大学管理学院;教育部过程优化与智能决策重点实验室;
【基金】：国家自然科学基金(71271071,71301041);国家自然科学基金重点项目(71490725) 国家“863”云制造主题项目(2015AA042101)~~
【分类号】：TP18;X513
【正文快照】： 1引言近年来,随着经济的快速发展,人类活动的加剧,土壤尘、燃煤、生物质燃烧、汽车尾气与垃圾焚烧、工业污染等因素使我国大部分地区遭遇持续性的雾霾天气,空气质量的不断恶化,雾霾天气现象频繁增多,对城市交通等各行各业及人类健康都造成严重影响.首先对于交通运输,受雾霾天

【相似文献】