Research On The Related Technologies In Joint-Radio-Resource-Management Of Heterogenous Wireless Networks

With the rapid development of Information and Telecommunication Technologies, different kinds of wireless networks emerge, taking on a heterogeneous network environment. The Joint radio resource management is proposed to achieve integrity, cooperation, and complementation among different RATs. Besides, the MBB services in future extend the traditional services from service number, service pattern, and also the requirement of network bearing, which beyond the capability of any single wireless network. From another point, the lack of adaption to the service behavior model results in the excessive deployment of base station equipment and excessive cost. Therefore, the application of heterogeneous networks to achieve resource allocation and artificial intelligence power-saving scheduling are discussed in this article.Based on the previous two aspects, the heterogeneous wireless network deployment scene, and the simulation platform compatible with multi-services are constructed in this article, which proposed the following schemes.In the aspect of bandwidth allocation in heterogeneous, a dynamic bandwidth allocation scheme for multiple services is proposed, and selecting the scene of traffic splitting. A bandwidth allocation which synthesizes both the service experience and network performance, in which access fairness of edge users in large scale wireless is proposed innovatively. In the procedure of algorithm, AHP is modified to adapt to different states of network load level and balance the weight between the protection of service experience and improvement of network performance, also the Dynamic thresholds and the utility function are applied to further guarantee the service fairness. A static and a dynamic simulation are applied to test the algorithm performance, and the simulation results shows that besides the ability to reduce the block rate of edge users remarkably, the heterogeneous wireless network performance is improved, i.e., keeping the block rate of the whole system sustain persistently low.In the aspect of power-saving scheduling in heterogeneous environment, different levels of learning agents are proposed to achieve a distributed learning system, and the signaling and information between the agents are defined and also the agent invoking scheme and interworking mode. Based on the system, Q-learning algorithm is applied to learn and adapt to user service behavior. To guarantee the service QoS in the procedure of power-saving, Q-learning algorithm is modified by introducing the prediction model in the action selection module, and annual intelligence neural network is applied to increase the generalization capability of machine learning. The simulation procedure are applied to test four performance, i.e. the effect of learning and decision, the performance of power-saving by the proposed scheme, the capability of guaranteeing the service QoS, and the details of power-saving procedure. The simulation shows that the proposed architecture of learning system well fit the power-saving scheme, and achieve efficient power-saving while guarantee the user QoS. Therefore, the proposed scheme achieve the purpose of design.Finally, a summarization of the content in this article is stated. The future research aspect of this article is proposed to achieve a further improvement.