| The future wireless network is a complicated heterogeneous network based on kinds of radio access technologies, which are different from each other and have their own advantages. Its target is to make communications with high quality wherever and whenever.The new wireless networks environments make the traditional technologies in homogeneous networks unavailable because of the complicated networks'application scenarios and the diversity of target users, so we need to make network selection. Moreover, due to the criteria of quality among many types of wireless networks and the requirements of quality among different people, the network selection in heterogeneous network becomes very difficult.To meet the requirement of growing business and diverse service provided by different networks, it is impractical to build only one kind of wireless network to satisfy the requirements above. So it is very important to study the technology of network selection algorithms.As one of the key technologies of heterogeneous network convergence, network selection should embody both the networks'advantages and the requirement of people.This paper presents a new network selection algorithm based on multi-objectives optimization, aiming for the end-to-end optimization, in which a utility function based on multi-objectives optimization is proposed. Multi-objectives such as end-to-end delay, available bandwidth and payoff are considered together.Based on this, this paper also presents a network selection scheme based on cognitive method, aiming at the development tendency of cognitive network. By the method of Q-learning, the scheme can predict the variation of the complicated environments.In the last part, simulation of the algorithm is designed by MATLAB to show the advantages or disadvantages. By comparing the new and the traditional algorithm, the simulation results show that the new algorithm has a large improvement in both the resource utilization and the load balance. |
PDF Full Text Request