摘要
鉴于传统的供应链求解方法所使用的线性模型考虑的影响因素具有局限性,不能适应互联网多源大数据环境下的动态供应链网络求解方法,将动态供应链映射为网络图模型,提出电子商务供应链网络模型,基于该模型,给出一种基于协同矩阵分解的半实例模式检测方法,用于在供应链网络中检测一个半实例化的协同行为模式。根据给定的协同行为模式,首先采用协同矩阵分解方法计算个性化供应链候选集;通过计算供应链网络节点实体之间的亲密度,运用A*图搜索算法,基于个性化供应链候选集生成供应链结果候选链集;根据个性化的时间、成本等其他限制条件,经过裁剪形成最后的供应链方案。通过面向服装领域的电子商务供应链数据集验证了方法的正确性、效率和准确性。
Owing to the fact that the consideration of linear model used by traditional supply chain solving method was limited,which was not met the requirement of dynamic supply chain network under Internet multi-source big data environment,the dynamic supply chain was mapped into graph model,and a half-instance model detection method based on collaborative matrix factorization was proposed on this basis,which was applied to detect a half-instantiated cooperative behavior on the dynamic supply chain network.According to a given collaborative behavior pattern,a personalized supply chain candidate set was computed with collaborative matrix factorization;by calculating the closeness degree between entity in supply chain network,apersonalized supply chain candidate set was formed with A＊algorithm graph search;according to other personalized restrictions such as time and cost,the final supply chain solutions was formed after clipping.Through the experiment of e-commerce supply chain data collection oriented to apparel field,the correctness,efficiency and accuracy of the method was verified.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2016年第2期324-329,共6页
Computer Integrated Manufacturing Systems
基金
国家自然科学基金资助项目(61572295
61272241)
创新方法工作专项资助项目(2015IM010200)
山东省自然科学基金资助项目(ZR2014FM031)
山东省自主创新重大专项资助项目(2015ZDXX0201B03
2015ZDXX0201A04
2015ZDJQ01002)
山东省重点研发计划资助项目(2015GGX101015
2015GGX101007
2014GGX101047
2014GGX101019)
山东大学基本科研业务费资助项目(2015JC031)
泰山产业领军人才工程专项经费资助项目~~
关键词
供应链
协同行为模式
A＊算法
supply chains
collaborative behavior patterns
A＊ algorithm
