摘要
基于统计与即时相结合的方法,设计了NGI中的一种切换目标预测机制.首先,采用统计预测,根据切换历史记录,建立模糊决策树,据此预测切换目标.如果预测到的目标有多个,则根据用户当前位置、运动方向和运动速率等即时信息以及平均呼叫持续时间进行即时预测,计算用户对每个预测到的目标的访问概率,选取该值最大的作为切换目标.仿真结果表明,同单一的统计预测机制和即时预测机制相比,该机制在不同速率下的预测命中率都较高,总体性能好.
Based on the combination of instant method with statistical method, a target prediction scheme for handoff in NGI is designed by use of statistical prediction and historical handoff record to form a fuzzy decision tree. If there exist several target cells predicted, an instant prediction scheme runs to compute the accessing probability of each and all predicted target cells in accordance to the user's current information such as position, moving speed/direction and average call duration, then the maximum addressing probability is picked out as the handoff target cell. Simulation results showed that the proposed scheme has higher hitting rates at different users' movement speeds in comparison with those schemes simply based on statistical prediction or instant prediction, with a better overall performance provided.
出处
《东北大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2007年第7期957-960,共4页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金资助项目(60673159)
国家高技术研究发展计划项目(2006AA01Z214)
新世纪优秀人才支持计划项目
高等学校博士学科点专项科研基金资助项目
辽宁省自然科学基金资助项目(20062022)
国家发改委CNGI示范工程项目(CNGI-04-13-2TCNGI-04-6-2TCNGI-04-15-7A)
关键词
下一代互联网
切换目标
统计预测
即时预测
模糊决策树
NGI ( next-generation Intemet )
handoff target
statistical prediction
instant prediction
fuzzy decision tree
作者简介
王兴伟(1968-),男,辽宁盖州人,东北大学教授,博士生导师;Correspondent: WANG Xing-wei, professor, E-mail: wangxw@ mail.neu. edu. cn
黄敏(1968-),女,福建长乐人,东北大学教授,博士生导师.