摘要
在原P-Grid构建算法的基础上,针对节点无初始数据负载量(Ignore-of-Load)及有初始数据负载量(Care-of-Load)两种情况提出改进的构建算法.对Ignore-of-Load算法,从加大路径延长的程度以及推荐成功率两方面提高其收敛速度;同时,提出了以路径为主导、以数据为主导及具有符合度调整的3种Care-of-Load算法,并进行了比较.模拟实验结果表明:Ignore-of-Load算法能减少原算法交互次数的50%以上,大幅提高了算法的收敛速度;具有符合度调整的Care-of-Load算法在收敛速度上表现良好,并且对数据索引的查找成功率在90%左右.
Based on the original P-Grid construction algorithm,two improved algorithms are proposed focusing on two situations,one is the node without initial data load（Ignore-of-Load） and the other is the node with initial data load（Care-of-Load）.The proposed Ignore-of-Load algorithm improves convergence rate in the aspects of extending the path with more bits and increasing the success rate of recommendation.The three Care-of-Load algorithms,two of which respectively focus on the path and the data and another of which possesses satisfaction adjustment,are then compared.Simulated results show that the proposed Ignore-of-Load algorithm reduces number of communications by more than 50 percentage points and greatly improves the convergence rate,and that the Care-of-Load algorithm with satisfaction adjustment can also improve the convergence rate and can search data index with a success probability of about 90%.
出处
《华南理工大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2010年第11期96-103,共8页
Journal of South China University of Technology(Natural Science Edition)
基金
教育部博士点基金资助项目(200805610019)
广东工业大学青年基金资助项目(082018)
