改进灰色模型在物流需求预测中的应用

Logistics Demand Forecasting Based on Improved Grey model

研究物流需求问题,物流受多种因素的综合影响,需求具有趋势性、较大波动性和随机性等变化特点,传统单一预测方法难以对其进行准确预测,为提高物流需求预测准确率,将灰色理论(GM)和支持向量机(SVM)相结合建立一种物流需求预测模型(GM-SVM)。GM-SVM首先采用灰色GM(1,1)预测模型动态预测物流需求变化趋势,然后运用SVM对GM(1,1)预测结果进行修正,以提高物流需求预测精度。采用具体物流需求实例对GM-SVM性能进行测试,实验结果表明,GM-SVM利用SVM和GM(1,1)的优势,达到优势互补,提高了物流需求的预测精度,更能全面描述物流需求的复杂变化规律。

Logistics demand is influenced by many factors, and has the characteristics of tend pattern, greater fluctuation and randomicity. The paper combined grey theory （GM） with support vector machine （SVM） to establish a logistics demand forecast model （ GM - SVM）. The GM - SVM first used gray GM （ 1,1 ） prediction model to fore- cast the dynamic change trend of logistics demand. Then SVM was used to correct the prediction results in order to improve the logistics demand forecast accuracy. The specific logistics demand examples were used to test the GM - SVM performances. The experimental results show that the GM - SVM makes use of the advantages of both SVM and GM（ 1,1 ） to achieve the complementary, which can improve the logistics demand forecast accuracy and describe more comprehensively the complex and nonlinear variation rules in logistics demands.