基于Q学习的异构网络选择新算法

针对异构网络中的多样业务需求，并且为了能够适应网络环境的动态变化，为每一个会话选择一个最合适的网络为其服务同时实现网络负载的均衡，以HSDPA和WiMax构成的异构网络为背景，基于Q学习算法，提出了一种异构网络环境下无线接入网络选择的新算法。该算法在进行网络选择时不仅考虑到网络的负载情况，还充分考虑了发起会话的业务属性、终端的移动性以及终端在网络中所处位置的不同。仿真结果表明该算法降低了系统阻塞率，提高了频谱效用，实现了网络选择的自主性。

New Q-learning based heterogeneous network selection algorithm

To meet the diverse service requirements in heterogeneous network and adapt to the dynamic changes in the network environment, choose the most suitable network for every session and realize network load balance, a new algorithm based on Q-learning for the selection of wireless access network in the heterogeneous network composed of HSDPA and WiMax was proposed. By this algorithm, the most suitable network for each session can be selected according to network load conditions, service attribute, mobility and different locations of terminal. At last, the simulation results show that the algorithm can lower system blocking probability, increase spectrum utility and realize the autonomy of network selection.