Research On Game Theoretic Resource Allocation Strategy In Wireless Networks

Game Theory studies how selfish individual compete and cooperate with each other. Two types of games are the non-cooperative and the cooperative game. Nash Equilibrium (NE), the solution of the non-cooperative game, is to achieve an equilibrium state of the competition between selfish individuals aiming at maxizing their own profits. As for the Nash Bargaining Solution (NBS), the commonly used solution of the cooperative game, is to maximize the whole collectivity's profits. NBS provides a binding agreement about the final resource allocation results among all candidates in the game. Thus, NBS enforces the fairness and generally has the Pareto optimum property. However, the NE usually needs the Pareto improvement. This paper uses the basic ideas and the classic mathematic models of games to analyze the problems of resource allocation in the cooperation transmission network and a hierarchical PMP/Mesh network. The main studies are summarized as follows.Firstly, for the resource allocation in frequency division multiple access (FDMA) based cooperative transmission networks, pricing based non-cooperative game is performed in two different models: (1) the monopoly market competition where multiple relays compete with each other to sell their spectrum resource to a single user, (2) the oligopoly market competition where multiple user nodes compete for the scarce spectrum resource from a single relay node. Both the benefits of the relay and the user are jointly maximized. The existence and the Pareto efficiency of the NE are proved. A distributed self-adaptive algorithm with its convergence conditions for solving the NE is given. Furthermore, a general method to improve the NE and to find a Pareto dominant solution that maximizes all the candidates'profits is provided. Simulation results show that the proposed game can stimulate cooperation of the relay node and coordinate the resource allocation reasonably. Besides, in the oligopoly scenario, a particle swarm optimization (PSO) algorithm is also developed as a searching method for the solution of the game. The performance comparisons under two different cooperative protocols, which are the amplify and forward (AF) and the decoded and forward (DF), are given by simulations.Secondly, a cooperative game is performed to solve the joint allocation of the number of cooperatively transmitted symbols and the symbol transmission power for time division multiple access (TDMA) based cooperative communication networks. In the game, a selfish node is energy and power limited, and is willing to seek for cooperative relaying only if the data-rate achieved through cooperation is not lower than that achieved through non-cooperation. To achieve the NBS, the game is divided into a symbol allocation game and a power allocation game. Simulation results show that the NBS is fair in that both nodes could experience better performance through cooperation, and also the NBS is efficient in that the performance loss of the NBS scheme is small.Thirdly, in order to provide higher capacity, throughput and QoS guarantee to territorial users in emergency scenarios, a low-medium altitude platform-based WiMAX system, AirWiMAX, is presented in this paper. A hierarchical PMP/Mesh topology is used, where an up-layer backbone network is constructed with the mesh structure and a PMP access network is supported in the down-layer. Then, a joint radio resource allocation is carried out simultaneously at the time, frequency and power domain for the AirWiMAX downlink. This problem is also modeled as a cooperative game with a fairness criterion. Simulation results show that compared with the other two typical resource allocation algorithms, i.e., the max-rate algorithm and the max-min fairness algorithm, the proposed algorithm achieves a good tradeoff between the overall system throughput and the fairness.Finally, based on the hierarchical PMP/Mesh topology, power control between Mesh client and Mesh BS in the uplink transmission is considered. In the code division multiple access (CDMA) system, each client aims to maximize its own transmission data bits under the condition of limited energy. Thus, a non-cooperative power control game is carried out. And the properties of the NE of the game are studied. Furthermore, the cost function is introduced to improve the NE with the supermodel game being applied to analyze the performance of the modified game. Simulation results show the efficiency of the power control algorithm.