OPT-UFR algorithm for distributed environment
COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(12B) 202-208
North China Institute of Aerospace Engineering, Langfang, Hebei, China
As a novel relevance filtering method, some experiments are done to illustrate the performance of UFR since this algorithm is proposed. However, there is no any comparison between UFR and some other relevance filtering mechanisms. This paper compares UFR with VON in scalability and efficiency. By changing the peer number and changing the AOI when setting the peer number fixed in experiments, it is proved that UFR is more efficient than VON in these two cases. Then some experiments on group based moving model prove the result more sensitive. To counter that the original “strip” algorithm to calculate the UFR border is not very efficient, we proposed OPT-UFR to reduce the useless update messages, and finally proposed a new algorithm to solve the heavy traffic problem for the joining node. Experimental results show that OPT-UFR always has better performance than both random based and group based moving models.