Research On Resource Allocation Algorithms In Femtocell Networks Based On Game Theory

Users raise demands for higher data rate and better QoS in the next generation of wireless network, most of which mainly come from indoor users. Due to the high cost of MBS, femtocell occurs as a supplementary technology to improve indoor coverage. Be characterized as low- power, low-cost and small coverage, femtocell improves indoor coverage as well as increasing system throughput and frequency usage. However, it brings interference to original cellular network and has become an urgent issue. This thesis explores resource allocation algorithms in femtocell networks based on game theory:(1) A power control algorithm related with price mechanism is researched. Under spectrum-sharing manner, the macrocell and femtocell together form a Stackelberg game. MBS sets an interference price for FUEs so as to inhibit the transmitting power of FUEs and protect itself from interference, then FUE determines its transmitting power according to the price, finally achieving a SE. Both sparse deployment scenario and dense deployment scenario are analyzed. Non-uniform price mechanism and uniform price mechanism are discussed separately.(2) In order to raise the performance of MUE, a joint power control and access control algorithm is proposed. A Stackelberg game is constructed, in which FAPs are leaders and MUEs are followers. FAPs firstly carry on a power control sub-game, aiming at optimizing the throughput of FUEs; then MUEs conduct an access control sub-game with the aim of optimizing the throughput of MUEs. Each MUE chooses a BS to optimize its performance and feeds back to FAPs. The game process will end till it reaches a stable equilibrium. Simulation results prove that this algorithm improves system throughput, especially for those MUEs neighboring FAPs.(3) A sub-channel allocation algorithm based on user classification and potential game is proposed. Cross-tier interference between macrocell and femtocell is considered with high priority. Given the cross-tier interference, MUEs are classified into different sets and an interference matrix is built, referred to which each FAP chooses its available sub-channels. After the first stage of channel allocation, a potential function is created. FAPs play a game and achieve their corresponding strategies. Because the set of available strategies decrease after the first-round of channel allocation, the complexity of this algorithm decreases to a large extent. Simulation results show that this algorithm is efficient either from the perspective of time complexity, or from the view of performance of users.