Cross-Layer Resource Allocation Design For Multi-User OFDM Systems

The Multiuser OFDM systems have the advantages of overcoming multipath fading,generating multi-user diversity easily, allocating resource flexibly, which can improve the capacityof the system. However, with the increase of the users’ QoS requirements, the traditionalhierarchical resource allocation policy has been not able to satisfy it, and the correspondingcross-layer resource allocation strategy due to combining multiple parameters for optimization, canmeet the users’ QoS perfectly.Firstly, the related knowledge of PHY and MAC layer for multi-user OFDM system isintroduced in this thesis. And then, according to the rate adaptive resource allocation criteria, someclassic dynamic resource algorithms are analyzed. Finally, a cross-layer dynamic subcarrierallocation algorithm is researched and simulated. However, because the algorithm doesn’t fully usethe rate limit, it restrains the improvement of users’ QoS.In this thesis, jointing PHY and MAC layer, a cross-layer resource allocation algorithm basedon the rate constraints is proposed in a multi-user OFDM system. According to analyzing the users’queue of MAC layer, the proposed algorithm finds the relationship between the packet loss and therate of physical, and uses it to get the users’ maximum and minimum rate constraints. In thephysical layer, firstly, with the cross-layer dynamic subcarrier allocation, the subcarriers areallocated initially. Secondly, according to the quality of channel state information, the remainingsubcarrier is allocated completely. Finally, through the users’ maximum rate constrains, users’ ratecan get further correction. The simulation shows that the algorithm can further reduce the users’packet loss, and through users’ maximum rate limit, some power can be saved. In addition, based onthe resource allocation model with the rate limit of the physical layer, a genetic algorithm isproposed to solution it in the thesis, in which the select operator uses an improved optimalpreservation strategy. The simulation results show that the capacity of the system can get a certainincrease, and more users’ QoS can be improved.