电商集中促销期间物流峰值统计模型仿真  

Simulation of Statistical Model of Logistics Peak Value in E-commerce's Centrally Sales Promotion Period

作　　者：王旭[1] 

机构地区：[1]南开大学经济学院,天津300071

出　　处：《计算机仿真》2015年第10期323-326,共4页Computer Simulation

摘　　要：电商集中促销期间会产生物流峰值,物流峰值的变化受到消费区域、经济水平、人员结构等复杂因素的影响,峰值出现的时间和大小的估计过程存在众多因素干扰。传统的估计方法在进行电商集中促销期间物流峰值统计的过程中,受到诸多印象的影响,物流峰值估计的相关参数选择缺少准确约束,使得物流峰值统计模型的可信度降低。提出基于K均值粒子群滤波的电商集中促销期间物流峰值统计方法。根据K均值算法对全部类型的电商进行分类,获取不同类型的电商对于物流峰值的影响程度。根据粒子群滤波算法的原理,将不同类型的电商看作一个个独立的粒子进行训练。根据训练结果,对当前全部电商集中促销期间的促销行为对物流峰值的影响进行估计。实验结果表明,改进算法能够对电商集中促销期间物流峰值进行精确统计。In the paper, a method for logistics peak value statistics in e - commerce＇s centrally sales promotion period based on K - means particle swarm filtering was proposed. According to K - means algorithm, all kinds of e - commerce were classified, to obtain the influence degree of different types of e - commerce on the logistics peak val- ue. According to the principle of particle swarm filtering algorithm, the different types of e - commerce were taken as a separate particle for training. According to the training results, the effect of all the current promotion behaviors in the e- commerce＇s centrally sales promotion period on logistics peak value was estimated. The experimental results show that the improved algorithm is able to make accurate logistics peak value statistics in e - commerce＇s centrally sales promotion period.

关 键 词：集中电商 促销 物流峰值 估计 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]