一种用于数据分布管理的模糊DR算法  

Fuzzy DR algorithm for data distribution management

作　　者：陈岭[1] 陈根才[1] 

机构地区：[1]浙江大学计算机科学与技术学院,浙江杭州310027

出　　处：《浙江大学学报（工学版）》2006年第9期1521-1525,共5页Journal of Zhejiang University：Engineering Science

基　　金：浙江省自然科学基金资助项目(Y104199);浙江省教育厅科研资助项目(20051020)

摘　　要：为了满足分布式交互仿真(DIS)中数据分布管理(DDM)对可扩展性和信息过滤的要求,提出一种模糊推算定位(DR)算法.该算法在多层DR算法基础上,选取系统中有用实体属性对,并确定其相应隶属函数.采用模糊一致关系从实体各属性间优先次序中得到其权值.通过模糊关联度计算在信息过滤过程中综合考虑实体属性,实现对任意两实体间相关性的衡量,计算结果被用来设置DR算法阈值.仿真结果表明,视角和实体大小属性均对信息过滤过程产生影响.由该算法过滤后的运动轨迹符合人的感知特点.准确过滤率数据显示,该算法可平衡系统可扩展性要求和精度要求.To meet the scalability and message filtering requirements of data distribution management （DDM） in distributed interactive simulation （DIS）, a fuzzy dead reckoning （DR） algorithm was proposed. Based on multi-level DR algorithm, useful entity attribute pairs of the system were chosen and corresponding subordinate functions were determined. Weights were distributed using fuzzy consistent relation for different entity attributes in their precedence orders. The attributes of entities were synthetically considered during the message filtering procedure by fuzzy correlation calculation, and the relativity between two arbitrary entities was measured. The threshold value of DR algorithm was determined upon the calculated value. Simulation results show that the view angle and the size attributes of entities affect the message filtering procedure, and that the movement trajectory filtered by the algorithm fits well the characteristics of human perception. Data on accurate filtering rate indicate that the algorithm can make a good tradeoff between the requirements of system scalability and accuracy.

关 键 词：分布交互仿真 数据分布管理 DR算法 数据过滤 模糊 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]