Power Allocation Of Multiuser Mimo Systems Based On Game Theory

Multiple-input multiple-output technique has been extensively developed during the last decade, however, the research on MIMO broadcast channel and MIMO interfering channel is little, the former and the latter corresponding to cellular system, WLAN respectively. In general, these two type of channels are the scenes of multiuser MIMO, under above two channel condition, how to utilize the communication resource effectively, such as frequency band, power is the important problem. The centralized optimal method is the best one in a view of spectrum efficiency, however, the computation complexity of it is too high. On the contrary, the distributed sub-optimal method realizes the good trade-off between spectrum efficiency and computation complexity. The noncooperative game theory is the classical representation of distributed method. So, this paper select it to optimize the mutual information of multiuser MIMO in a fully distributed fashion. The content of this paper is as follows:Firstly, as a typical example of multiuser MIMO, the power allocation of MIMO interference channel is considered. This allocation can be formulated as a problem of non-cooperative game theory. In the framework of game theory, Nash Equilibrium existence is proved. In order to get to the Nash Equilibrium, a distributed algorithm implemented iteratively is designed. Next, Nash Equilibrium uniqueness and the global convergence of asynchronous distributed algorithm are proved. Finally, simulation results show algorithm convergence.Secondly, the distributed resource allocation for D2 D communications underlaying cellular networks is considered. In cellular network, the D2 D links and cellular users can be regarded as multipoint to multipoint parallel communication systems and point to point virtual MIMO system, thus the channel among D2 D links and cellular user can be viewd as MIMO interference channel. Next, this special MIMO channel is formulated as non-cooperative game theory problem, and the most suitable strategy of every player is the waterfilling algorithm. Furthermore, the existence and uniqueness of Nash Equilibrium is proved. For validating the convergence of algorithm, some simulations are implemented.