基于移动Agent和最优搜索的分布式信息检索系统研究

发布时间：2018-04-17 16:01

本文选题：分布式信息检索 + 移动Agent　；参考：《电子科技大学》2013年硕士论文

【摘要】：随着计算机技术的发展和网络技术的普及，互联网已成为人们获取信息的最重要途径之一。搜索引擎技术虽然在一定程度上解决了网络上资源的定位问题，但是随着网络规模的扩大，特别是移动互联网的崛起，使得网络信息逐渐向异构、复杂和分布式特性发展，传统的信息检索技术越来越明显的体现出其局限性。移动Agent作为新一代分布式人工智能技术以其自主性、移动性和智能性等特点，克服了传统分布式计算模式的固有缺陷，为现有信息检索技术的发展提供了新的思路。虽然基于移动Agent的分布式信息检索技术给人们带来了一种新的检索模式，但是还存在一些问题制约着它的发展，如检索时间的限制，检索效果不理想等。因此，如何制定更好的检索策略成为目前亟待解决的问题。本文以此为出发点重点研究了在有限搜索时间条件下，提高信息检索系统搜索质量的方法。最优搜索理论是二战时期发展起来的一门学科，其解决的核心问题就是在搜索资源有限的情况下，寻找一种最优的分配方案，使得在有限资源条件下找到目标的概率达到最高，或使成功找到目标所消耗的资源最少。因此最优搜索理论很适合用于处理信息检索系统的性能优化问题。本文首先在研究了现有基于移动Agent的分布式信息检索技术，以及最优搜索相关理论的基础上，，建立了基于移动Agent的分布式信息检索优化模型。然后按照最优搜索理论的思想，提出了适合于分布式检索系统的最优时间分配模型和最小期望成本模型，并对初始概率分布的确定及其调整方法进行了深入研究，分析并确定了探测函数的形式，最后给出了在有限搜索时间限制下的最优搜索方案。本文还提出了一种对资源节点搜索质量的反馈评价方法，并将其应用到最优分配策略的自动更新上，因此增强了系统的自适应能力。针对节点网络环境的变化以及软硬件之间的差异，提出了一种移动Agent的迁移协商机制，实现了在资源检索过程中的Agent路由动态更新，从而在一定程度上实现了系统的负载均衡，有效节省了检索时间。另外，本文根据上面所建立的模型，设计并实现了基于Agent的分布式信息检索系统，最后在系统上进行了相关实验，结果表明，以上措施对于系统做到了切实的改进。
[Abstract]:With the development of computer technology and the popularization of network technology, the Internet has become one of the most important ways for people to obtain information.Search engine technology has solved the problem of resource location in the network to a certain extent, but with the expansion of network scale, especially the rise of mobile Internet, network information has gradually developed into heterogeneous, complex and distributed characteristics.The limitation of traditional information retrieval technology is more and more obvious.As a new generation of distributed artificial intelligence technology, mobile Agent overcomes the inherent defects of traditional distributed computing mode and provides a new way of thinking for the development of the existing information retrieval technology because of its autonomy, mobility and intelligence.Although the distributed information retrieval technology based on mobile Agent has brought us a new retrieval mode, there are still some problems restricting its development, such as the limitation of retrieval time, the unsatisfactory retrieval effect and so on.Therefore, how to formulate a better retrieval strategy has become an urgent problem.This paper focuses on improving the search quality of information retrieval system under the condition of limited search time.The theory of optimal search is a discipline developed in the period of World War II. The core problem it solves is to find an optimal allocation scheme under the condition of limited search resources, so that the probability of finding a target under the condition of limited resources is the highest.Or it takes the least amount of resources to succeed in finding a goal.Therefore, the optimal search theory is suitable for the performance optimization of information retrieval systems.Based on the research of the existing distributed information retrieval technology based on mobile Agent and the theory of optimal search, this paper establishes a distributed information retrieval optimization model based on mobile Agent.Then, according to the idea of optimal search theory, an optimal time allocation model and a minimum expected cost model are proposed for distributed retrieval systems, and the determination of initial probability distribution and its adjustment methods are deeply studied.The form of detection function is analyzed and determined. Finally, the optimal search scheme is given under the limit of finite search time.This paper also proposes a feedback evaluation method for the search quality of resource nodes, and applies it to the automatic updating of optimal allocation strategies, so that the adaptive ability of the system is enhanced.In view of the change of node network environment and the difference between hardware and software, a migration negotiation mechanism of mobile Agent is proposed, which realizes the dynamic update of Agent routing in the process of resource retrieval, and thus realizes the load balance of the system to a certain extent.The retrieval time is saved effectively.In addition, this paper designs and implements a distributed information retrieval system based on Agent according to the above model. Finally, some experiments are carried out on the system. The results show that the above measures can improve the system effectively.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.3

【参考文献】