基于产业链协同平台的配件跨链商品库构建及搜索系统研究

发布时间：2018-08-10 19:32

【摘要】：随着集群式供应链的发展,企业节约成本和提高效率的方法不再局限于企业自身的能力和资源,单供应链上的核心企业不仅需要整合链条上下游的资源,还需要关注不同单供应链之间跨链协作的竞合关系,以提高链条之间信息的交流和反馈市场的速度。配件供应是汽车产业链中的重要一环,不同用户角色均有掌握配件相关信息的需求,本文通过构建配件跨链商品库和实现搜索系统来解决现有配件管理系统中存在的问题。首先根据平台各企业数据库的物理结构特点设计了配件跨链商品库,并利用针对需要提高响应速度的配件库优化问题,对物化视图技术进行了研究,根据常用的统计查询语句,运用谱系聚集方法归纳查询集合,利用考虑了查询成本和维护成本的物化权重公式,选择合适的查询语句进行物化,用定时刷新物化视图保证了数据的准确性和有效性。针对用户搜索需求,搜集信息建立配件常用词汇表,采用基于同义词词林的词语相似度计算方法,对用户输入的关键词进行扩展,并运用基于全文检索和模式图的关系数据库关键词搜索方法,先通过全文索引找到包含用户输入关键词的元组集,再用关键词元组集扩展配件库模式图得到扩展模式图,并运用改进后的候选网络生成算法,融合词语相似度,提出对元组连接树的排序机制,将元组连接树按顺序转化为SQL语句,交由数据库执行,最终将结果以二维表的形式展示给用户。同时对改进后的候选网络生成算法和原算法进行实验,从查全率、查准率和F1值三方面对算法效果进行比较分析。本文以企业配件库中配件销售业务数据的分析程作为示例,利用谱系的聚集过程图对常用复杂查询语句进行归纳,创建了提高配件库响应速度的物化视图。系统使用SQLserver数据库和C#语言进行开发,采用基于B/S模式的三层架构,设计实现了基于产业链协同平台的配件搜索系统。
[Abstract]:With the development of cluster supply chain, the methods of saving cost and improving efficiency are no longer limited to the ability and resources of the enterprise itself. The core enterprises in the single supply chain need not only integrate the upstream and downstream resources of the chain. It is also necessary to pay attention to the competition and cooperation between different single supply chains in order to improve the speed of information exchange and feedback market. Spare parts supply is an important part in automobile industry chain. Different user roles have the need to master the relevant information of accessories. This paper solves the problems existing in the existing parts management system by constructing the cross-chain commodity library of accessories and realizing the search system. First of all, according to the physical structure of each enterprise database of the platform, the cross-chain commodity library of accessories is designed, and the materialized view technology is studied according to the optimization problem of spare parts library which needs to improve the response speed, and according to the commonly used statistical query statements, The method of pedigree aggregation is used to induce the query set and the materialized weight formula considering the query cost and maintenance cost is used to materialize the query statements. The accuracy and validity of the data are guaranteed by refreshing the materialized view periodically. According to the users' search needs, the common vocabulary of accessories is established by collecting information, and the words similarity calculation method based on synonym forest is adopted to expand the keywords input by users. Using the search method of relational database key words based on full-text retrieval and schema graph, we first find the tuple set containing user input keywords through full-text index, and then use keyword tuple set extension component library schema diagram to get the extended schema map. Using the improved candidate network generation algorithm, combining the word similarity, the sorting mechanism of the tuple join tree is proposed, and the tuple join tree is transformed into SQL statements in order, which is executed by the database. Finally, the results are presented to the user in the form of a two-dimensional table. At the same time, the improved candidate network generation algorithm and the original algorithm are experimented, and the results of the algorithm are compared and analyzed from three aspects: recall rate, precision rate and F1 value. This paper takes the analysis process of spare parts sales business data in the enterprise accessories library as an example, sums up the common complex query statements by using the aggregate process diagram of the pedigree, and creates a materialized view to improve the response speed of the accessories library. The system uses SQLserver database and C # language to develop, adopts the three-tier architecture based on B / S mode, designs and implements the accessory search system based on the industrial chain cooperation platform.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3;F274

【参考文献】