基于Multi-Agent模型结构的数字文献编目检索系统设计

Design of digital document cataloging and retrieval system based on Multi⁃Agent model structure

传统系统存在召回率和检索精准度低的问题,为此,文中提出基于Multi-Agent模型结构的数字文献编目检索系统设计。根据系统硬件结构,依次对用户界面层、功能模块及开发工具层、核心对象层和系统支撑层进行分析。其中,用户界面层向用户提供Web交互式图形界面;功能模块及开发工具层向用户提供开发工具,通过交互信息协同完成任务;核心对象层按照信息交互策略向功能模块提供所需数据;系统支撑层向系统底层数据对象提供维护和协调功能,维护数据一致性。以面向Multi-Agent模型的软件工程为设计理念,计算执行任务与数字文献编目任务处理数值,以现有检索技术为基础,设计基于Multi-Agent模型的检索流程。由实验结果可知,该系统最高召回率和检索精准度都为95%,具有较高的实用性。

The traditional system has the defects of low recall rate and retrieval accuracy.Therefore,the design of a digital document cataloging retrieval system based on the Multi⁃Agent model structure is proposed.According to the hardware structure of the system,the user interface layer,functional module and development tool layer,core object layer and system support layer are analyzed in turn.The user interface layer provides users with a Web interactive graphical interface.The function module and development tool layer provide users with development tools to complete tasks by interactive information collaboration.The core object layer provides the required data to the function module according to the information interaction strategy.The system support layer provides maintenance and coordination functions to the underlying data objects of the system to keep the data's consistency.Multi⁃Agent model⁃oriented software engineering is used as the design concept to calculate the numerical values of the executed tasks and digital document cataloging tasks.On the basis of the existing retrieval technologies,a retrieval process based on the Multi⁃Agent model is designed.The experimental results show that both the maximum recall rate and retrieval accuracy of the system are all 95%.Therefore,the system has high practicality.