基于改进的思维进化算法与BP神经网络的AQI预测  

AQI Prediction Based on Improved Mind Evolutionary Algorithm and BP Neural Network

作　　者：高帅 胡红萍[1] 李洋 白艳萍[1] GAO Shuai;HU Hong-ping;LI Yang;BAI Yan-ping(School of Science,North University of China,Taiyuan 030051,China;School of Electronic Engineering,Xi＇an University of Electronic Science and Technology,Xi＇an 710071 China)

机构地区：[1]中北大学理学院,山西太原030051 [2]西安电子科技大学电子工程学院,陕西西安710071

出　　处：《数学的实践与认识》2018年第19期151-157,共7页Mathematics in Practice and Theory

基　　金：国家自然科学基金(61774137);山西省自然科学基金(201701D121012,201701D221121)

摘　　要：思维进化算法（MEA）的趋同和异化操作带有太多的随机性，公告板的信息不能得到充分利用，使得效率下降，出现重复搜索．为了避免MEA算法的这个缺点，借鉴遗传算法（GA）和粒子群算法（PSO）的优点，提出改进的思维进化算法（MEA—PSO—GA）．利用MEA—PSO—GA算法优化BP神经网络的初始权值和阈值进而预测太原的日常空气质量指数（AQI）．通过与MEA—BP算法，MEA—PSO—BP算法和MEA—GA—BP算法比较，实验结果表明，提出的MEA—PSO—GA—BP算法在预测精度、误差率和可靠性方面搜索速度更优，更易于实现AQI预测，具有较好的有效性和可行性，有一定的现实意义．The convergent and alienated operations of Mind Evolutionary Algorithm （MEA） have too much randomness, and bulletin board information can not be fully used, resulting in the decrease of efficiency and repeated search. In order to avoid this shortcoming of MEA algorithm, this paper proposes an improved thought evolutionary algorithm （MEA-PSO-GA） based on the advantages of Genetic Algorithm （GA） and Particle Swarm Optimization （PSO）. In this paper, we use MEA-PSO-GA to optimize the initial weights and thresholds of BP neural networks to predict the daily air quality index （AQI） of Talyuan. Compared with MEA-BP algorithm, MEA-PSO-BP algorithm and MEA-GA-BP algorithm, the experimental results show that MEA-PSO-GA-BP algorithm proposed in this paper has more search speed in prediction accuracy, error rate and reliability Excellent, easier to achieve AQI prediction, has good effectiveness and feasibility, and has certain practical significance.

关 键 词：思维进化算法 粒子群优化算法 遗传算法 BP神经网络 空气质量预测 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]