Study On The Heterogeneous Wireless Network Selection Algorithm Of Supporting QoS

With the development of wireless network,many kinds of networks with different structure came into sight,such as WLAN,GSM,CDMA,WiMAX.It makes that users can choose different service supplied by different network operators according to their special requirement.Networks with different structure differ in many aspects,that makes the wireless more and more complicated.And with the development of users' equipment,many kinds of transactions appear in the users' equipment,and different transaction should be connected with different network.So when people enjoy the benefit of the technologies' development,they also encountered a difficult problem,it is that how to ensure that they can always connect to the network which is most suitable to them,so the algorithm of choosing wireless network appears.This algorithm is to ensure that users can connect to a most suitable network when the connected network can't satisfy users' requirement.Previous network selection algorithm is based on a single attribute of wireless network for selecting,but in the next generation network,it will have huge error if the algorithm still adopts single standard because of the complexity is increased,so the algorithm should adopt many network performance evaluation,it is also known as multiple attribute decision making(MADM).It proposes two algorithms based on MADM in this paper by comparing other algorithms,they are Wireless network selection algorithm based on independence of influence factors of QoS and wireless network selection algorithm based on QoS objectivity.Wireless network selection algorithm based on independence of influence factors of QoS comply with the principle of the lower nodes are the reason of the upper nodes strictly,this removes the interdependence between those network's properties effectively.Wireless network selection algorithm based on QoS objectivity uses the method that more than one person make the decision matrix and calculate the average,thenverify the correctness of selection result according to the feedback.This algorithm reduces the error of weight of all these network's properties effectively.We can get a conclusion that these two algorithms are better than others in aspects of network selection and error control through the experiments and data analysis.