Research And Simulation Implementation Of Dynamic Energy-efficient Network Selection Algorithm In Heterogeneous Networks

With the rapid development of mobile wireless communication and mobile internet technology,the future wireless communication network will be integrated with various wireless access technologies like 4Q 3G,WLAN and WiMax which is a heterogeneous wireless network environment.What's more,Multi-mode terminal technology also allows users to make full use of network resources and make flexible and sophisticated services possible.However,how to design the access network selection mechanism to make the always best connection decisions while satisfying user's QoS requirements with reasonable allocation of network resources becomes a key problem within the heterogeneous network integration environment.At the same time,the terminal energy consumption problem has also become a limiting factor in the development of the applications and services.But,among the studies about heterogeneous network selection algorithm,most of which adopt multiple attribute decision method to make the network selection decision,energy consumption is rarely considered.So,it is of great significance to choose the best connected target while saving terminal energy.In this thesis,aiming at the network selection problem,we put forward a dynamic energy-efficient network selection algorithm.First,we study the energy efficient network selection algorithm based on fuzzy neural network.we calculate energy utility value of different network access ports according to different types of services.Then set the energy utility value,residual network bandwidth,received signal strength as well as user's movement speed as the input of the fuzzy neural network and next,make the optimal selection through the process of fuzzification,logic inference and defuzzification.By the way of setting difference of blocking rate of candidate networks as the inference goal in fuzzy neural network,the blocking rate of two networks tend to be equal,which makes the use of the heterogeneous network resources reasonably.our proposed fuzzy neural network based network selection algorithm considers the port energy consumption and other QoS parameters,which greatly reduces the energy consumption of the terminal during the statistics time and provides high quality service.Second,pointing at the problems of the dynamic change of network QoS parameters with uncertainty,this thesis studies Q-learning based dynamic network energy-efficient network selection algorithm.We calculate energy consumption value of different network access ports according to different types of services.Then,we consider user QoE,handover cost between networks and terminal energy consumption as the reward function of Q-learning method.Based on the goal of maximum accumulated reward,the terminal interacts with the changing network parameters and makes the choice dynamically.Our algorithm gains a higher quality of the user experience with lower energy consumption in different cases of service distribution,which achieves the goal of saving energy.By comparing the simulation results,the proposed two algorithms can dynamically adapt to the changes of the network parameters and select the best network which indicates the superiority in energy saving.