Abstract: Logistics quantity forecasting is a key link of logistics system planning and logistics resource allocating. It is also a fundamental work to formulate development plan and policies for the logistics industry; therefore, it is rather important to forecast logistics quantity accurately. At present, the single forecast method is commonly adopted to forecast logistics quantity, which has its limits in assumed conditions and adaptation and always leads to low accuracy. Based on the multiple linear regression method, second-index flatness method, and fixed weight combination forecast method, the study proposes the weight changeable combination method to forecast the logistics quantity. An example conducted for freight volumes forest in Liling city of Hunan province suggest that the proposed method is more accurate and more rational to forecast logistics quantity than the single forecast method.

Key words: logistics quantity, single forecast method, weight changeable combination forecast method, freight volume

摘要： 物流量预测是物流系统规划、物流资源合理配置过程中的重要环节，同时也是政府部门制定物流产业发展规划与政策的基础工作，准确预测物流量具有非常重要的意义。目前，大多数学者往往都采用单一预测方法进行预测，但由于单一预测方法假设条件及适用范围均存在一定的局限性，预测精度不高。因此，在多元线性回归方法、二次指数平滑方法和定权组合预测方法的基础上，提出了一种物流量的变权组合预测方法，并以湖南省醴陵市货运量预测为实例进行了分析，结果表明,该方法较单一预测方法误差更小、精度更高，能够更准确地预测物流量，可以作为物流量预测的有效工具。

关键词: 物流量, 单一预测方法, 变权组合预测, 货运量