Research On Fast Cell Association Scheme In Ultra Dense Cellular Network

With the increasing demand of high-speed data services and the rapid development of mobile communication technology,the mobile data traffic and the number of User Equipment(UE)in the fifth generation mobile communication system(5G)will increase exponentially.In order to deal with the above problems,the Ultra Dense Cellular Network(UDCN)which densely deploys small cells has been applied to the5 G communication network,although the intensive feature of UDCN improves the system capacity of the mobile communication network to a certain extent,it also results in a single small cell serving only one or less UEs,which not only leads to the waste of network resources but also the poor fairness of UE.It is well known that various cell association schemes can effectively improve the performance of cellular networks,but existing schemes mainly optimize a single objective and could not transform such a problem into a polynomial time solvable 2-dimensional matching problem with a generic utility function.Therefore in UDCN network scenario,we propose a general objective function polynomial time cell association scheme,which transforms the cell association problem into a 2-dimensional assignment problem solvable by Hungarian algorithm.Specifically,we get an experience pool to store multiple performance indicators of cellular network based on this scheme.Then we obtain the results of cell association corresponding to the optimization of the given performance indicators in the cell association decision-making stage.At the same time,compared with the existing cell association schemes,this scheme achieves a better trade-off between the improvement of cellular network system capacity and UE fairness and reaches lower computational time.Although the aforementionedcell association scheme acquired a good trade-off of cellular network performance,with the rapid increase of the number of small cells and UEs in the communication network,in order to achieve low latency of 5G communication network,we need to design a cell association scheme that can improve the overall performance of the cellular network and obtain lower computational time further.Therefore,with the goal of maximizing the proportional fairness of cellular network,we further propose a Hungarian Algorithm Based Back Propagation Neural Network(HA-BPNN)cell association scheme.First,we transform the proportional fairness cell association problem into a two-dimensional matching problem that can be solved by the Hungarian algorithm.Second,the cell association result obtained by the Hungarian algorithm and corresponding user equipment spectrum efficiency matrix will be used to complete the training of BPNN.At last,not only does the well trained BPNN achieve approximate optimization of the proportional fairness of cellular network,but also it obtains better cellular network system capacity.At the same time it greatly reduces the computational time.