基于贝叶斯网络模型的物流成本预测与控制研究  

Research on Logistics Cost Prediction and Control Based on Bayesian Network Model

作　　者：董寅华 王成义[2] Dong Yinhua;Wang Chengyi(Shanghai Polytechnic of Communications,Shanghai 200431,China;Zhejiang Sci-Tech University,Hangzhou 311121,China)

机构地区：[1]上海交通职业技术学院,上海200431 [2]浙江理工大学,浙江杭州311121

出　　处：《现代科学仪器》2024年第6期375-380,385,共7页Modern Scientific Instruments

基　　金：浙江省科技厅科学技术支持项目(编号:21ZJ837405)。

摘　　要：准确预测物流成本是提高物流配送效率的关键,而传统物流企业缺乏对物流运输成本的预测,进一步导致物流成本增加,企业营收收入低。为进一步提高物流成本预测与控制,提出贝叶斯网络模型,并结合局部加权回归方法建立物流成本预测模型。将物流成本特征之间的变量进行分类,利用最大似然估计提高贝叶斯网络模型的成本数据分类效率。实验结果表明,使用贝叶斯网络模型的物流成本预测值与实际成本值的曲线误差较小,最小误差值仅为0.68%,实际运输成本与预测成本差值仅为1元,进一步表明贝叶斯网络模型可以非常准确地预测物流成本。且预测评估指标中的平均绝对百分比误差与均方根误差均呈下降趋势,最小数值分别为2.21%、3.62。贝叶斯网络模型预测模型的置信度设置为80%-85%,具有较高的成本预测可靠性。Accurately predicting logistics costs is the key to improving logistics distribution efficiency,while traditional logistics enterprises lack prediction of logistics transportation costs,which further leads to an increase in logistics costs and low revenue for enterprises.To further improve logistics cost prediction and control,a Bayesian network model is proposed,and a logistics cost prediction model is established by combining local weighted regression methods.Classify the variables between logistics cost characteristics and use maximum likelihood estimation to improve the cost data classification efficiency of Bayesian network models.The experimental results show that the curve error between the logistics cost prediction value and the actual cost value using the Bayesian network model is small,with a minimum error value of only 0.68%.The difference between the actual transportation cost and the predicted cost is only 1 yuan,further indicating that the Bayesian network model can accurately predict logistics costs.The average absolute percentage error and root mean square error in the predictive evaluation indicators are both decreasing,with the minimum values being 2.21%and 3.62,respectively.The confidence level of the Bayesian network model prediction model is set to 80%-85%,which has high cost prediction reliability.

关 键 词：贝叶斯网络模型 物流运输成本 预测 局部加权回归 置信度 

分 类 号：F252[经济管理—国民经济]