杨建成

摘 要： 物流需求是多种因素综合作用的结果，具有规律性和特殊性，变分十分复杂，导致当前物流需求预测模型的预测效果差，难以满足物流管理的实际应用要求。为了解决物流需求建模过程中存在的难题，提出基于ARIMA?SVM的物流需求预测模型。首先对当前物流需求预测的研究现状进行分析，找到引起物流需求预测效果的原因;然后选择差分自回归滑动平均模型对物流需求的规律性变化特点进行建模，支持向量机对物流需求的特殊性变化特点进行建模;最后采用权值方式确定物流需求预测的预测结果，并采用物流需求的预测实例分析模型的有效性。结果表明，ARIMA?SVM的物流需求预测结果要优于当前其他物流需求预测模型，为其他预测问题提供了一种建模工具。

关键词： 物流管理; 随机性变化特点; ARIMA?SVM; 权值的确定; 预测模型; 支持向量机

中图分类号： TN911.1?34; TP391 文献标识码： A 文章编号： 1004?373X（2018）09?0182?05

Abstract： The logistics demand affected by many factors has the characteristics of regularity， particularity and complex variation， which leads to the poor prediction effect of current logistics demand forecasting model， and is difficult to meet the practical application requirements of logistics management. In order to solve the problems existing in logistics demand modeling process， a logistics demand prediction model based on ARIMA?SVM is proposed. The research status of the current logistics demand forecasting is analyzed to find out the reason influencing the logistics demand forecasting results. The autoregressive integrated moving average （ARIMA） model is selected to model the regular variation characteristics of logistics demand. The support vector machine （SVM） is used to model the special variation characteristics of logistics demand. The weight is used to determine the prediction results of the logistics demand forecasting. The validity of the logistics demand forecasting model is analyzed by means of an example. The results show that the result of logistics demand forecasting model based on ARIMA?SVM is better than that of other logistics demand forecasting models， which provides a modeling tool for other forecasting problems.

Keywords： logistics management; stochastic variation characteristic; ARIMA?SVM; weight determination; prediction model; support vector machine

0 引 言

随着经济全球化进展的不断加速，物流行业得到了前所未有的发展，而物流园区规划是物流业发展的基础[1]。物流需求的建模与预测是合理规划物流园区的前提条件。物流需求的建模与预测是指对某区域的物流历史数据和资料进行收集和分析，对将来物流需求进行估计，以描述物流需求将来的变化趋势[2]。物流需求的建模与预测涉及多个方面的技术，如计算机技术、信息处理技术、数量统计技术等，因此物流需求的建模与预测面临着巨大的挑战[3?5]。……