基于移动Agent的分布式搜索应用研究

发布时间：2018-07-25 13:23

【摘要】：近年来，随着互联网的广泛应用，互联网信息呈现爆炸式增长。海量资源的分布性、开放性，，狭窄的、不稳定的信息传输渠道，导致人们难以准确、快速的获取有价值的信息。针对这种情况，一种在现有硬件条件下，用软件解决，既能够有效运用网络带宽，又可以高效处理网络资源的移动Agent技术应运而生。它集软件、通讯、分布式技术于一体，有着巨大的前景。在分布式信息搜索方面，移动Agent技术可以将代码迁移到资源所在地，对本地资源建立索引并进行搜索运算，最后携带用户的检索结果返回检索请求端，很好的解决了传统搜索引擎搜索异地资源速度慢、系统开销高的不足。本文首先介绍了当前国内外信息检索技术的现状。针对当前分布式搜索技术的不足，引入了移动Agent技术。在对移动Agent技术进行研究的基础上，提出了一个基于个性化的移动Agent信息搜索模型使移动Agent技术与分类算法的结合，并对移动Agent迁移的路由选择问题及并行计算涉及到的移动Agent之间的通信技术、容错功能和负载均衡等关键技术进行了研究。论文根据中文信息搜索的特殊性，在原有的最大匹配分词算法的基础上进行改进，引入了用户搜索的预处理过程。在路由选择上，回避了传统的利用网络状况及传输距离等因素为选择条件的不确定情况，而采用基于内容的用户兴趣度的路由选择策略。最后在IBM的Aglet移动Agent开发工具下，分析该模型与传统的信息搜索的优势，证明了该模型的适用范围，为高效化的信息搜索研究奠定了基础。
[Abstract]:In recent years, with the wide application of the Internet, the Internet information has an explosive growth. The distribution, openness, narrow and unstable information transmission channels of massive resources make it difficult for people to obtain valuable information accurately and quickly. In order to solve this problem, a mobile Agent technology, which can use network bandwidth effectively and deal with network resources efficiently, has emerged under the existing hardware conditions and solved by software. It integrates software, communication and distributed technology, and has a great prospect. In the aspect of distributed information search, mobile Agent technology can migrate the code to the location of the resource, index and search the local resource, and finally bring the user's search results back to the request side. A good solution to the traditional search engine search for resources in remote areas slow speed, high system overhead. This paper first introduces the current situation of information retrieval technology at home and abroad. Aiming at the deficiency of current distributed search technology, mobile Agent technology is introduced. Based on the research of mobile Agent technology, a personalized mobile Agent information search model is proposed to combine mobile Agent technology with classification algorithm. The routing problem of mobile Agent migration and the key technologies of communication between mobile Agent, fault-tolerant function and load balancing involved in parallel computing are also studied. According to the particularity of Chinese information search, this paper improves on the original algorithm of maximum matching word segmentation, and introduces the preprocessing process of user search. In route selection, the traditional route selection strategy based on user interest based on content is used to avoid the traditional uncertainty of the selection condition, such as network condition and transmission distance. Finally, the advantages of the model and the traditional information search are analyzed under the Aglet mobile Agent development tool of IBM, and the application scope of the model is proved, which lays a foundation for the research of high efficiency information search.
【学位授予单位】：华南理工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP391.3;TP18

【相似文献】