Efficient dynamic resource allocations for OFDM systems

Abstract

In this dissertation, we study on the efficient resource (bit/power/subcarrier) allocation for OFDM systems. To reduce the computational complexity of an allocation algorithm, we propose three new schemes; 1) Dimension reduction method, where group-based allocation algorithm is exploited, 2) Reduction of modulation modes, which is achieved by the strict mathematical modelling and proofs, and 3) Fast Lagrangian techniques, where the computational complexity is reduced in both the iterative optimal algorithm and the non-iterative suboptimal algorithm. For the resource allocation, the channel information at receiver side is fed back to a transmitter with a certain time delay. The performance improvement by the resource allocation can be degraded due to this feedback delay and time-varying characteristics of wireless channel. Two different approaches are proposed to compensate the delayed channel feedback. The first approach is to obtain frequency diversity over selected multiple subbands. The proposed schemes have both benefits of diversity transmission and subband allocation. In the second approach, the probability that BER is over target BER is kept constant regardless of mobile speeds by controlling the transmission power. Total transmission power is minimized by efficiently allocating subbands to users.