Energy Management And Resource Allocation In Green Relay-Assisted LTE-A Networks

With the rapid development of Internet Technology and the growth of mobile terminal number,mobile data traffic is experiencing explosive growth,which greatly increases the energy consumption of the wireless network.Thus,in recent years,the concern on how to reduce the energy consumption in wireless communication is increasing.The introducing of renewable energy mitigates the adverse impact on traditional energy resources and the environment on the one hand,and facilitates the network deployment of poor power supply area on the other hand.Further,as the QoS(Quality of Service)requirement from users is increasing,new networking solutions management strategy is in urgent need.For these two points,green stations powered by renewable energy show great advantages,especially for the traffic sparse or poor electricity supply area: firstly,by utilizing green stations to unload the traffic of macro base station,the grid power consumption of the macro base station will decrease;secondly,introducing green stations as relays will improve the quality of communication link and thus increase the data rate.However,the time-varying characteristic of the energy harvesting brings new challenges to the hybrid networking and energy management of the heterogeneous network.We consider the green relay-assisted network,and propose different algorithms in the following three scenarios:1.Green station is introduced to offload the traffic of the macro station,and the energy saving of the grid power consumption is optimized;2.Deploying green stations in the edge of the macro station's coverage,energy saving and quality of service are both taken into consideration;3.Green stations are introduced into the D2 D communication system to assist the transmission,thus system capacity is improved.For the first scenario,green station sleep and user association is studied.The state of green stations and the user association in each time slot is optimized to maximize the energy saving in the long term.Since the time scale of green station sleeping and user association are different,thus the original optimization problem is divided into two sub-problems.Inner and outer cross iteration loop is used to obtain the optimal solution.The simulation results demonstrate the significant effect of the deployment of green stations on network energy saving and the effectiveness of the developed algorithm.For the second scenario,dynamic green station sleeping mechanism is studied.The mismath between energy harvesting and traffic arrival is mitigated by dynamically adjust the state of green stations in the time scale,thus the long term energy consumption is reduced.Dynamic programming approach is adopted to divide the original optimization problem into sub-problems in different time slots,and solution is obtained by performing backward induction of the costto-go function.Because of the computational complexity of the joint optimization of all green stations and based on the independence of the decision of each green station,reduced DP algorithm is developed,which decides the per-RS sleep ratio iteratively.By simulations,the reduced DP algorithm achieves satisfactory performance and outperforms the greedy algorithm.For the third scenario,the joint optimization of resource allocation and green relay selection is studied in D2 D communication system.It consists of two cross-iteration parts: green relay selection algorithm based on the survival probability,and the resource allocation and power control algorithm in LTE-A.In the green relay selection algorithm,by introducing the concept of survival probability,the available green relay set is obtained at the beginning of each frame by estimating the approximate probability of each green relay's survival probability.Then the transmission assisted relay is selected by each user among the available green relay set.In the resource allocation and power control algorithm,by relaxing the binary variable and introducing a new auxiliary variable the mixed-integer non-linear problem is transformed into a convex optimization problem.Then Lagrange dual method is adopted to obtain the resource allocation to each user and the power allocation on each resource block.Since the survival probability is related to the energy consumption,thus after the resource and power allocation,the survival probability and the available green relay set will be refreshed.The algorithm is proved to be effective by the simulations.