摘要
为提升城市货运量需求预测精度,构建了PMGM(1,N)灰色马尔可夫预测模型,并以宁波市和芜湖市为例进行实证研究。首先,建立多因素指标体系,并通过灰色关联分析厘清核心变量;然后,采用粒子群算法(Particle swarm optimization,PSO)优化MGM(1,N)模型的背景值插值系数,提升模型拟合能力;最后,引入马尔可夫状态转移矩阵修正残差,提高模型对数据波动的适应性。实证结果表明,PMGM(1,N)灰色马尔可夫模型在拟合精度和预测精度上均优于GM(1,1)、多元线性回归、MGM(1,N)及PMGM(1,N)模型,充分验证了其在城市货运需求预测中的有效性与优越性。
To improve the accuracy of urban freight demand forecasting,a PMGM(1,N)grey Mar⁃kov prediction model was developed,with Ningbo City and Wuhu City as empirical study cases.First,a multi-factor indicator system was established,and grey relational analysis was employed to identify the core variables.Then,the particle swarm optimization(PSO)algorithm was applied to optimize the background value interpolation coefficient of the MGM(1,N)model,enhancing its fitting capability.Fi⁃nally,a Markov state transition matrix was introduced to correct the residuals,thereby improving the model's adaptability to data fluctuations.The empirical results show that the PMGM(1,N)grey Mar⁃kov model outperforms GM(1,1),multiple linear regression,MGM(1,N),and PMGM(1,N)models in both fitting and forecasting accuracy,fully verifying its effectiveness and superiority in urban freight demand forecasting.
作者
刘星辉
任建伟
李雪琦
LIU Xinghui;REN Jianwei;LI Xueqi(Institute of Transportation,Inner Mongolia University,Hohhot 010070,China)
出处
《内蒙古大学学报(自然科学版)》
2025年第6期656-668,共13页
Journal of Inner Mongolia University:Natural Science Edition
基金
国家自然科学基金项目(72262024)。
