期刊
GREY SYSTEMS-THEORY AND APPLICATION
卷 11, 期 3, 页码 498-517出版社
EMERALD GROUP PUBLISHING LTD
DOI: 10.1108/GS-02-2020-0023
关键词
Grey system model; Grey wolf optimizer; PM10; Euler polynomial; Air pollution; Euler polynomial-driven grey model
资金
- National Natural Science Foundation of China [71901184, 71771033, 71571157]
- Humanities and Social Science Fund of Ministry of Education of China [19YJCZH119]
- Sichuan Province Undergraduate Training Programs for Innovation and Entrepreneurship [S201910619005S]
- Grey System Theme Innovation Zone [GS2019017]
- National Statistical Scientific Research Project [2018LY42]
This study introduces the use of grey system theory to build a short-term PM10 forecasting model, incorporating the Euler polynomial and grey wolf optimizer to improve flexibility and accuracy. With high performance in the short-term PM10 forecasting case in Tianjin China, the model shows potential in accurately forecasting PM10 concentration in big cities of China, serving as a decision-making support tool in the future.
Purpose PM10 is one of the most dangerous air pollutants which is harmful to the ecological system and human health. Accurate forecasting of PM10 concentration makes it easier for the government to make efficient decisions and policies. However, the PM10 concentration, particularly, the emerging short-term concentration has high uncertainties as it is often impacted by many factors and also time varying. Above all, a new methodology which can overcome such difficulties is needed. Design/methodology/approach The grey system theory is used to build the short-term PM10 forecasting model. The Euler polynomial is used as a driving term of the proposed grey model, and then the convolutional solution is applied to make the new model computationally feasible. The grey wolf optimizer is used to select the optimal nonlinear parameters of the proposed model. Findings The introduction of the Euler polynomial makes the new model more flexible and more general as it can yield several other conventional grey models under certain conditions. The new model presents significantly higher performance, is more accurate and also more stable, than the six existing grey models in three real-world cases and the case of short-term PM10 forecasting in Tianjin China. Practical implications With high performance in the real-world case in Tianjin China, the proposed model appears to have high potential to accurately forecast the PM10 concentration in big cities of China. Therefore, it can be considered as a decision-making support tool in the near future. Originality/value This is the first work introducing the Euler polynomial to the grey system models, and a more general formulation of existing grey models is also obtained. The modelling pattern used in this paper can be used as an example for building other similar nonlinear grey models. The practical example of short-term PM10 forecasting in Tianjin China is also presented for the first time.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据