A GA-Based Optimization for Frequency-Selective Scheduling in Downlink OFDMA Systems with Cyclic-Delay Diversity

Abstract

CDD (cyclic-delay diversity) aided frequency-selective scheduling has been known as an effective technique to increase system capacity of the OFDMA (orthogonal frequency-division multiple access) systems. By increasing the channels' frequency selectivity with cyclic delays, the merit of multi-user diversity can be exploited more effectively in the system. In this paper, we aim to optimize the design of the CDD-aided frequency-selective scheduling for downlink OFDMA systems. A genetic algorithm (GA)-based optimization is devised to obtain the cyclic-delay values which provides a higher system sum-rate than the multi-degree adaptive cyclic-delay diversity method and has a lower computational complexity than the exhaustive method with only a slight degradation in system sum-rate. The effectiveness of the proposed method is verified by extensive computer simulations.