基于会话特征的MAP选择算法

Session-character based MAP selection algorithm

针对现有MAP选择算法的评价指标不能直接反映用户的实际通信情况的问题,提出了新的指标——会话阻塞率.利用排队理论推导出会话阻塞率的数学期望公式,分析得到影响会话阻塞率的4个因素:会话频率、持续时间、节点的速度、MAP的层次.同时根据实时性需求和会话阻塞率间的关系,提出了基于会话特征的MAP选择算法.该算法采用了基于速度选择算法的节点分类机制,并在选择条件上增加了节点的部分会话特征.通过实验对这2种算法的性能进行比较和分析,结果表明本算法能够为用户的实时性需求提供更好的支持.

Since the current evaluation guidelines for mobile anchor point （MAP） selection algorithms can not manifest the actual condition of user＇s communication, a new guideline, session blocking rate （Tint ） , is proposed. According to the queuing theory, the derived mathematical expectation of Tint, is related to four factors： session ratio, session duration, the speed of mobile node and the layer of MAP. In order to adjust the Tint to the user＇s requirement for real time communication, session-character based MAP selection algorithm is proposed. This algorithm adopts the node classification mechanism, which is used in speed-based selection algorithm, and adds some session characters to the selecting conditions. From the numeric results, it is concluded that session-character based MAP selection algorithm can select the proper MAP according to user＇s requirement for real time communication.