基于多属性决策的异构网络选择算法比较研究

本文提出层次分析法结合逼近理想解法共同组成的算法对备选网络进行网络选择,主要考虑带宽、时延、抖动和丢包率四个QoS属性以及价格属性对网络选择的影响,判断了四类典型业务下,网络的综合性能,最后选择出适合用户业务的最佳网络。仿真结果表明,该算法适用于流类业务较多的情况,与层次分析法相比,该算法既能准确选择适合用户业务的接入网络,同时还能将网络负载从16.61%下降到了8.63%。

Comparative Studies on Multiple Attribute Decision Making of Heterogeneous Network Selection Algorithm

The proposed algorithm combines AHP with TOPSIS to select optimal access network. Focusing on the influences of QoS parameters to network selection, we take five attributes into consideration: bandwidth, delay, jitter, packet loss probability and cost. And then comprehensive performances of candidate networks with four different services are decided. Finally, the Always-Suitable-Connection for users is selected. The simulation results show that, comparing to the AHP, the proposed algorithm fits for the case with more streaming service. It not only can accurately select the optimal access network with different services, but also it can reduce the network load from 16.61% to 8.63% at the same time.