Context Information Based Wireless Resource Allocation In Cellular Network

Along with the rapid spread of intelligent terminals and high-speed development of the mobile-Internet,users' service requirements and communication scenarios are increasingly diversified,and the mobile data shows explosive growth.In response to the rapid growth of massive data and to achieve support for the new scenes,the fifth generation mobile communication system proposed a target aiming at achieving 1000 times growth in system capacity,and technologies such as massive multiple input multiple output,ultra dense network and device to device communication(D2D)are regarded as key technologies of 5G communication system.On the other hand,while increasing the processing overhead,the rapid increase of data also brought the potential value.By excavating the implicit information of the context information in particular scenarios,the user behaviors can be analyzed to predict their future requests,and it is of significant importance to apply such results to resource allocation in new scenarios of 5G to further improve system capacity.In cellular network,according to the 5G communication system technology scenarios,scenes can be divided into continuous wide area coverage,hot spot high capacity,low delay high reliability and low power consumption with large connection,etc.Due to the difference of service requirements as well as the acquisition and utilization of the context information in different scenes,their impact on the resource allocation in cellular network is different.Therefore,the dissertation will focus on dense small cell networks and D2D communications underlaying cellular networks,and research the context information based wireless resource allocation strategies in such two specific scenarios.For dense small cell networks,due to the limited capacity of the backhaul in the scenario,it is common to consider deploying a cache on the small cell to avoid duplication of popular files.Thus the transmission rate of user depends on not only physical channel condition but also hit ratio of the cache and capacity of the backhaul.Based on such background,we initialize the cache content through user based Top N collaborative filtering recommender system according to the context information like user history request,and then propose a resource allocation strategy maximizing system throughput under the proportional user throughput constraint.Since the formulated problem is non-convex and involves variables including user association and cache determination,it is difficult to solve it directly,thus we propose a user association updating algorithm and a cache content updating algorithm to iterate and update these two variables,and finally allocate spectrum resource under the proportional user throughput constraint.Simulation results reveal the obvious gain in hit-ratio and system throughput compared to the existed algorithms.For D2D communications underlaying cellular networks,on account of the scarcity of spectrum resource,D2D pairs have to share the spectrum resource of cellular users which results in interference between D2D pairs and cellular users and reduces the service quality of cellular users.However,people tend to share resource with close friends in real lives.Based on such background,we take advantage of the principle and interactive information in social network to define and calculate user closeness among users.And then we jointly consider physical channel condition and user closeness and propose a resource allocation scheme to maximize system throughput while satisfying the quality of service requirement of both cellular users and D2D pairs.Since the optimization problem is Mixed Integer Non Linear Programming(MINLP),which is difficult to obtain the optimal solution,we decompose the optimization problem into three subproblems:admission control,power allocation,and optimal CU-D2D pair matching and resolve them one by one.Simulation results show that the proposed algorithm is superior to the four comparing algorithms in terms of system throughput,which is better suited for such spectrum shortage scenarios.