基于Multi-Agent协同进化机制的Web个性化信息推荐的研究  

The Research on Web Personalized Information Recommendation based on Multi-Agent Co-evolutionary Mechanism

作　　者：易晨晖 Yi Chenhui(Hunan Institute of Information Technology,Changsha Hunan 410151,China)

机构地区：[1]湖南信息学院,湖南长沙410151

出　　处：《信息与电脑》2017年第14期94-96,99,共4页Information & Computer

基　　金：2017年湖南省教育厅科学研究项目"基于Multi-Agent的关联规则挖掘在Web访问行为上个性化推荐的研究"(项目编号:17C1115)

摘　　要：Web个性化信息推荐是近年来的研究热点,但是面对Web信息量大、数据异构及信息安全等难题,以及信息推荐中信息过滤系统不透明、算法歧视、信息窄化等新问题,笔者根据Multi-Agent协同进化机制具有的并行性和自组织、自适应、自学习等智能特征,提出将Multi-Agent协同进化机制应用于Web个性化信息推荐,设计了基于Multi-Agent协同进化机制的Web个性化信息推荐系统。其中Multi-Agent根据内外部环境变化对适应度进行检测和评估而进行协同进化,由最优Agent产生推荐信息。这样有效帮助用户准确、快速地获得Web个性化推荐信息。Web personalized information recommendation is research focus in recent years.However,in the face of such challenges as enormous web information,data heterogeneity and information security,as well as new problems such as opacity of information filtering system,algorithm discrimination and information narrowing in information recommendation,considering the advantages of the intelligent characteristics of Multi-Agent co-evolutionary mechanism,such as parallel and self-organization,self-adaptive,self-learning and so on,this paper proposes the application of Multi-Agent co-evolutionary mechanism to Web personalized information recommendation,the Web personalization information recommendation system based on multi-agent collaborative evolution mechanism was designed.The Multi-Agent fitness is tested and evaluated according to internal and external environment changes and then Multi-Agent evolves,recommendation information is generated by the optimal Agent.It effectively helps users get the Web personalized recommendation information accurately and quickly.

关 键 词：MULTI-AGENT 协同进化机制 Web个性化信息推荐 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]