摘要
针对经济社会系统中广泛存在的时滞因果问题,通过分析各驱动因素对主系统行为的时滞累积作用以及各驱动因素序列之间存在的非线性关系,构建了基于动态分析的时滞累积多变量灰色ATDGPM(1,N)预测模型,并探讨其参数求解方法。基于动态分析的时滞灰关联向量确定了驱动因素和时滞期;利用粒子群优化算法对幂指数进行优化求解;论证了DGM(1,N),DGPM(1,N)和ATDGM(1,N)模型均是该模型在不同参数取值下的特殊形式。数值实验结果表明ATDGPM(1,N)模型能够更好的描述系统行为序列与驱动因素序列之间的时滞非线性关系,从而有效提高建模精度。将该模型应用于河南省粮食产量的模拟和预测中,可得ATDGPM(1,N)模型的模拟和预测精度远远高于DGPM(1,N)模型和GM(1,N)模型,从而进一步验证了模型的有效性和可行性。
As a large agricultural province in China,Henan province holds an important position in the overall situation of national food security,and the forecast of grain production is an important research topic in agricultural economics.This study forecasts the grain production in Henan province and puts forward some suggestions from the perspective of the government’s policy making.Aiming at the widespread time-lag causality problem in economic and social systems,by analyzing the time-lag cumulative effect of each driving factor on the behavior of the main system and the nonlinear relationship between the sequences of each driving factor,we construct a time-lag cumulative multivariate grey ATDGPM(1,N)prediction model based on the dynamics analysis,and explore its parameter solving method.The application scope of the grey multivariate prediction model is broadened from the theoretical method,and some suggestions are provided for the decision-making of government departments from the practical point of view.Based on the data of China Statistical Yearbook and Henan Provincial Statistical Yearbook from 2007 to 2020,this paper constructs a time-lag cumulative multivariate grey ATDGPM(1,N)prediction model based on dynamic analysis,and proposes the“Dynamic Grey Correlation Time-Lag Analysis Method”to determine the time-lag relationship between the sequence of the system behaviors and that of the related factors.By analyzing the time lag relationship between the system behavior sequence and the related factor one,the dynamic grey correlation time lag analysis method is proposed to determine the driving factors and the time lag period;from the perspective of improving the accuracy of the model,with the objective of minimizing the average error of modeling,the known conditions such as the model parameters and the time-response equation are taken as the constraints,the nonlinear optimization model is constructed,and the particle swarm optimization algorithm is utilized for the optimization of the power index.It is argued that the DGM(1,N),DGPM(1,N),and ATDGM(1,N)models are all special forms of the model for different values of the parameters.The results of numerical experiments show that the ATDGPM(1,N)model can better describe the time-lagged nonlinear relationship between the sequence of system behaviors and that of driving factors,so as to effectively improve the modeling accuracy,and the numerical experiments show that the prediction accuracy of the model constructed in this paper is higher than existing literature;the model is applied to the simulation and prediction of grain output in Henan province,and the DGM(1,N)model predicts the grain output of Henan province from 2021 to 2020 through the DGM(1,N)model.The analysis results show that among the main input factors of grain output in Henan province from 2021 to 2024,the drought-affected area and fertilizer application show a trend of a decrease year by year,and the total power of agricultural machinery and the effective irrigated area show a trend of an increase year by year.In 2021,the grain output in Henan province was reduced by the impact of floods.Other than that,the grain output of Henan province in the next three years maintains the trend of continuous growth.This development trend can be seen through the model’s prediction results,which indicate that Henan province will gradually move from a large agricultural province to a strong agricultural one,continue to face the country’s major needs,and give full play to its geographic advantages,so that the grain production can increase and stabilize,and contribute wisdom and strength to ensuring food security.Then from the perspective of the government to formulate policies,(1)we should improve the farmland water conservancy and irrigation system,increase investment in water conservancy,and complete large and medium-sized irrigation areas with water-saving renovation and large and medium-sized irrigation and drainage pumping station upgrading;(2)we need to improve the level of agricultural mechanization,strengthen the guidance of agricultural machinery,increase the number of reserve arable land,improve the ability to cope with climate change and natural disasters,and focus on scientific and technological initiatives;(3)we ought to improve the support policy and increase support.It is recommended that government departments and society at large should actively unblock agricultural supply channels,promote the development of the agricultural economy,further stabilize policy support,and focus on preventing and resolving major risks.
作者
罗党
马艳
LUO Dang;MA Yan(School of Mathematics and Statistics,North China University of Water Resources and Electric Power,Zhengzhou 450046,China)
出处
《运筹与管理》
CSSCI
CSCD
北大核心
2024年第12期195-202,共8页
Operations Research and Management Science
基金
国家自然科学基金资助项目(51979106)
华北水利水电大学研究生创新项目(YK2021-113)。
关键词
动态分析
时滞累积
ATDGPM(1
N)
粒子群算法
粮食产量预测
dynamic analysis
time-delayed accumulative
the ATDGPM(1,N)
the particle swarm optimization algorithm
grain yield forecasting
