Dynamic Bound Restriction Iterative Algorithm Framework and Application to Power and Bit Allocation in OFDM Systems

Abstract

This paper proposes a dynamic bound restriction iterative algorithm (DBRIA) framework, which can make iterative algorithms converge quickly. We apply this framework to realize computationally efficient iteration-based suboptimal bit and power allocation in OFDM systems. The implementation methods for three classes of suboptimal allocation algorithms are analyzed; margin-based, water-filling-based, and Lagrange-multiplier-based. We extend the water-filling-based power allocation algorithm to implement the bit allocation as well. Additional techniques for further reducing the computational complexity for all three allocation methods are also discussed. Simulation results show that both the maximum and the average computational complexity of all three allocation algorithms are reduced dramatically by applying the DBRIA and the additional techniques. For the water-filling-based algorithm, the allocation performance is also improved.