基于无偏灰色PSOMarkov优化模型的加油站便利店零售销量预测方法研究

A New Prediction Method on Retail Sales of Convenience Stores in Gas Stations Based on Unbiased Grey Markov Optimization Model

加油站便利店是一种新兴的零售业态,准确地预测其零售销量具有重要意义.以往的销量预测模型多为单一模型,提出了一种基于粒子群优化的无偏灰色PSOMarkov预测方法.首先,基于灰色系统理论建立了无偏GM(1,1)预测模型,消除了常规GM(1,1)预测模型的固有偏差.其次,利用Markov理论对无偏GM(1,1)预测模型的相对残差进行了修正,模型能较好地体现数据的波动特征.最后,利用改进的粒子群优化算法白化无偏GM(1,1)-Markov预测模型灰区间的参数,得到无偏灰色PSO-Markov预测模型.通过云南昆明市红瓦副加油站便利店的零售销量案例表明,模型能提高预测模型的精度.模型可用于加油站便利店的商品销售预测,并为企业的经营决策提供依据.

The convenience store of gas station is a new retail formats in China.It is important to predict its retail sales accurately.The sales forecasting models are mostly single models at previous research.In this paper,an unbiased grey Markov prediction method based on particle swarm optimization(PSO) is proposed.Firstly,an unbiased gray GM(1,1) forecasting model is established based on the grey system theory,which can eliminate the inherent bias of the original gray prediction model.Then,according to the theory knowledge of Markov chains,an unbiased GM(1,1)-Markov model is built to correct relative residuals of unbiased GM(1,1) model,which can reflect the volatility characteristics of the data.Finally,based on a new algorithm of particle swarm optimization(PSO),an optimization of unbiased GM(1,1)-Markov model is set up to albino the parameters of gray interval.The unbiased GM(1,1)-Markov-PSO model is obtained.A forecast case on retail sales of convenience store of gas station in Kunming is shown,which improves the prediction accuracy of the model significantly.Thus,the model can be used to predict the actual sales of convenience store,and it can provide the basis for decision-making of enterprise.