主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2025, Vol. 33 ›› Issue (12): 185-199.doi: 10.16381/j.cnki.issn1003-207x.2023.1297cstr: 32146.14.j.cnki.issn1003-207x.2023.1297

• • 上一篇    下一篇

欧勒多项式驱动的自适应多变量灰色预测模型及应用

丁圆苹, 党耀国, 王俊杰()   

  1. 南京航空航天大学经济与管理学院,江苏 南京 211106
  • 收稿日期:2023-08-14 修回日期:2023-12-02 出版日期:2025-12-25 发布日期:2025-12-25
  • 通讯作者: 王俊杰 E-mail:wangjj@nuaa.edu.cn
  • 基金资助:
    国家自然科学基金项目(72001107);国家自然科学基金项目(72271120);国家自然科学基金项目(71771119);国家自然科学基金项目(72571136);教育部人文社会科学研究青年基金项目(19YJC630167);中国博士后科学基金项目(2020T130297);中国博士后科学基金项目(2019M660119);江苏省自然科学基金青年项目(BK20190426);中央高校基本科研业务费项目(NP2022104);江苏省研究生科研与实践创新计划项目(KYCX24_0508);中国国家留学基金项目(202406830002)

Adaptive Multivariable Grey Prediction Model Driven by Euler Polynomials and Its Application

Yuanping Ding, Yaoguo Dang, Junjie Wang()   

  1. School of Economics and Management,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China
  • Received:2023-08-14 Revised:2023-12-02 Online:2025-12-25 Published:2025-12-25
  • Contact: Junjie Wang E-mail:wangjj@nuaa.edu.cn

摘要:

不同时间尺度下的能源环境系统数据具有不同的变化趋势特征。针对具有线性、非线性、或部分线性部分非线性混合变化趋势特征的数据序列预测问题,构建含有非线性时间扰动参数的欧勒多项式,灵活表征数据序列的复杂变化趋势,进而提出欧勒多项式驱动的自适应多变量灰色预测模型。结合灰导数信息覆盖原理及欧勒多项式均值公式,推导出模型的差分形式和离散形式,并给出基于灰狼优化算法的欧勒多项式阶数和非线性时间扰动参数寻优框架。在此基础上,利用数学归纳法求解时间响应式。最后,选取具有不同变化趋势特征的北京市月度PM2.5和江苏省年度能源消费总量实例验证模型的有效性,进而将模型应用于广州市全社会季度用电量预测。同时,基于蒙特卡洛仿真分析以及不同比例的训练集建模,验证新模型的鲁棒性与稳定性,说明新构建的多变量灰色预测模型能够自适应拟合和预测具有不同变化趋势特征的数据序列。

关键词: 灰色预测, 欧勒多项式, GM(1, N)模型, 用电量预测

Abstract:

The data of energy environment system under different time scales have different trend characteristics. Aiming at the prediction of data series with linear, nonlinear, or partially linear, partially nonlinear mixed trends, the Euler polynomial with nonlinear time disturbance parameters is constructed to flexibly characterize the complex trends of the data series, and then the adaptive multivariable grey prediction model driven by Euler polynomial is proposed. The difference and discrete forms of the model are derived based on the grey derivative information covering principle and the Euler polynomial mean value formula, and the optimization framework of Euler polynomial order and nonlinear time disturbance parameters based on the grey Wolf optimization algorithm is given. On this basis, the time response formula is solved by mathematical induction. Finally, examples of monthly PM2.5 in Beijing and annual total energy consumption in Jiangsu Province with different trend change characteristics are selected to verify the validity of the model, and then the model is applied to the quarterly electricity consumption forecast of Guangzhou. Meanwhile, the robustness and stability of the new model are verified based on Monte-Carlo simulation analysis and different proportions of training set modeling, indicating that the newly constructed multivariate grey prediction model can adaptively fit and predict data series with different trend change characteristics.

Key words: grey prediction, Euler polynomial, GM(1, N) model, electricity consumption forecast

中图分类号: