摘要
考虑国内连锁零售业需求预测的现状,以一家典型经营快速消费品的连锁零售食品公司为研究对象,通过分析其现有的门店销售预测和总仓预测模式,借助订货公式找出需求时界,确定预测时段,采用动态优化加权移动平均,自适应指数平滑,综合预测等方法提高预测的准确性,同时检验现有经验模式的可行性。研究结果表明7天移动平均的方法非常适合长期预测,综合预测的方法可以提高预测准确性,但其实施的复杂程度增加。同一种预测方法在不同订货周期预测过程中,误差区别较大,门店销售受诸多因素影响,需求波动较大,不同订货周期的门店应采用不同的预测方法。
Considering the domestic present demand forecasting situation of the retail chain,a typical fast moving consumer goods-selling retail chain on food is taken as a research sample. Through the analysis of present pattern of forecasting in store and commodity center,find out the demand time fence and confirm the forecasting period based on the ordering formula to improve the accuracy of forecasting with usage of dynamic optimizing method of weighted moving average, adaptive exponential smoothing model,integrated forecasting and etc.The feasibility of the present empirical pattern is examined at the same time.This conclusion illustrates the seven-day weighted moving average method is more appropriate for the long period forecasting comparing to the integrated forecasting for the implementation and accuracy.It also indicates the error difference is significant between different ordering period forecasting and the store sales is affected by a lot factors which contribute to the demand fluctuation.Therefore,different forecasting methods should be taken for different ordering period stores.
出处
《工业工程与管理》
CSSCI
北大核心
2014年第3期60-65,71,共7页
Industrial Engineering and Management
基金
国家自然科学基金资助项目(71072035)
关键词
连锁零售
预测
周期
retail chain
forecasting
period
