Research On User Association Techniques In Dense Wireless Network

In recent years,with the development of data services and the popularization of intelligence terminal devices,the demand of mobile data is growing rapidly,which has put forward a lot of challenges to the current wireless communication systems.And along with the advancement of the fifth generation mobile communication system(5G),a novel dense network technology has become the key technology and research trend to meet the rapidly growing user requirements for its high spectrum efficiency,high channel capacity,high transmit rates,self-organization and low latency through deploying a large number of small access points in cellular networks densely.And the visible light communication after 5G is also an ideal realization scheme for indoor hot spot communication with the help of dense wireless network technology.To expand the system capacity by a big margin,the denser network deployment multiplexes the spectrum resources,but it also causes serious frequency interference,which affects the quality of service for users.Therefor,a rational resource allocation and user scheduling technology has been the key factor that affects the efficiency of dense network.At the same time,to evaluate the rationality of user scheduling technology,it often requires a survey of multiple performance indicators and the user fairness.In this connection,two-side matching model between users and resource nodes is built as the theoretical basis for the analysis of user association problems,and based on this,the problem of joint optimization of multiple utilities and global optimization of one utility in dense wireless network user association are the main research contents.First of all,joint optimization of user rate and fairness in small cell dense heterogeneous network is considered.We simplify the user association problem as the many-to-many matching problem for users and resources,and then subdivide the resource nodes so that the many-to-many matching probelm turns into the many-to-one matching model for users and subcarrier,and use the monotonicity analysis of bilateral utility function to solve the externalities in communication scenario,which the matching theory is applied.Fur-thermore,two distributed algorithm based on the deferred acceptance algorithm is put forward and we further show that the proposed method can achieve both user-wise and system-wise optimality.The simulation results indicate that the proposed algorthims outperform existing user association approaches in terns of throughput and fairness,and can convergence quickly to a user scheduling scheme which takes both efficiency and fair-ness into account.In the sequence,spectrum resource allocation in D2D dense dense heterogeneous network for achieving global optimization is considered.With the help of local interaction game theoretic,the global maximum of thoughput and the global minimum of collision level are changed into the local altruistic game and local con-gestion game.Afterwards,two local interaction games are proved as exact potential games,which makes the problem of achieving global optimization a optimal Nash equilibrium problem.Two distributied algorithms based on spatial adaptive play are proposed to achieve the global optimum.The simulation shows that both algorithms can converge to the global optimum and the second algorthm is more suitable for D2D scenario with high network density for its convergence rate positively related to network density.Finally,we consider the user association problem of indoor visible light dense cellular network.To achieve the maximization of system utilities,we change the target from only one resource-user pair into multiple resource-user pairs based on two-side exchange stable matching,so that the many to one matching model between users and access points is established.We use the model to capture the externalities with utility functions,rather than traditional preference ordering,and then prove that local optimal of sum of utilities meets the characteristics of two-side exchange stable matching and the existence of exchange stable matching,and the maximization of system utilities is translated into the maximization of two-side exchange stable matchings.Based on learning mechanism,a algorithm for maximizing sum of utilities with adaptive adjustment of iterative speed is put forward.The simulation shows that the algorithm can converge to the global optimum quickly and do better than traditonal approaches in sum of user utilities.