Research On User Perception Downlink Scheduling Algorithm In LTE Networks

With the development of the wireless communication industry,the demand for multimedia data service transmission has exploded in recent years.And also wireless communication has gradually exposed many shortcomings,especially in terms of frequency spectrum utilization and data transmission rates.Since different types of services have different requirements on the transmission rate,there must be a scheduler in e Node B to select the appropriate channel resources to allocate to every user equipments(UEs)according user requirements.At the same time,the high-mobility of the cellular mobile network determines the diversity of the environment in e Node B coverage area.Therefore,in the long term evolution(LTE)system,designing a reasonable downlink scheduling algorithm for different complex environments is the key to ensuring wireless transmission and improving the user's experience with the network.For this purpose,this thesis will combine the practical application background to study and propose some new downlink intelligent scheduling algorithms,in order to solve the problem of wireless resource management under the condition of limited bandwidth resources.Moreover,schedulers should also provide higher quality of experience(Qo E)for wireless user terminals,while maintaining system performance.The main work of the thesis can be briefly summarized as the following two parts.First of all,aiming at the problem that some emergency services cannot be transmitted at a high priority in multi-service transmission,the widely used proportional fairness(PF)scheduling algorithm is firstly studied in depth,then the weights are added based on the definition of proportional fairness,and a series of theoretical deductions are made.Finally,an exponential weighted proportional fairness(EWPF)algorithm is proposed.Due to limitations of some indicators,a category of UEs with specific traffic types or other transmission characteristics want to be prioritized in transit.By setting the weights,the EWPF algorithm can solve that problem,and the user experience is also improved accordingly.The simulation results and the using of the introduced user perceived throughput show that the EWPF method can not only guarantee the system throughput and fairness,but also reduce the effective transmission time of users.Then the scheduling process under different scheduling strategies is compared and analyzed,which verifies the effectiveness of the algorithm.Secondly,for the problem that the scheduling algorithm with fixed parameters cannot adapt to the dynamic scene changing along with time,two algorithms,PPF-RL-1 and PPF-RL-2,are proposed.Based on one priority function which is related to the percentage of datatransferred,the two algorithms combine the idea of actor-critic(AC)algorithm belonging to reinforcement learning.This structure enables the base station to dynamically adjust the scheduling strategy of base station to cell users according to the change of user channel conditions or other factors.Finally,the simulation results show that these two algorithms can adapt to various complex environments and improve users' experience for the communication network,and also improve system performance while guaranteeing system fair index.