Chapter 1 – Network Technologies 39 Genetic-Based Subcarrier and Bit Allocation Algorithm for Multiuser OFDM System M.M.Elmesalawy, G.A.Elfadeel and I.I.Ibrahim Network Research Group, Faculty of Engineering, University of Helwan, Cairo, Egypt. e-mail: Mlf_egt@yahoo.com, Gam_hel@yahoo.com, Ibsmail@softhome.net Abstract This paper proposes an adaptive subcarrier and bit allocation technique for multiuser OFDM (Orthogonal Frequency Division Multiplexing) system based on genetic algorithm. The objective is to minimize the overall transmission power while ensuring the per user’s data rate and Bit Error Rate (BER) requirements. This has been achieved by developing a search algorithm based on genetic system. Simulation results show that the performance of the proposed algorithm outperforms the Wong’s adaptive OFDM (MAO) algorithm in terms of computational complexity. Keywords OFDM, Multiuser, Subcarrier, Bit, Allocation, Genetic. 1. Introduction It has been suggested that multiuser OFDM systems employing adaptive subcarrier and bit allocation can take advantage of the channel diversity among users in different locations, thereby enabling an efficient use of all subcarriers in the OFDM system. Optimal bit loading and subcarrier-allocations have already been formulated by, for example, Wong in (C. Y. Wong, et al., 1999) , (W. Rhee and J. M. Cioffi, 2002): more specifically, by minimizing the overall transmission power under a given data rate constraint in (C. Y. Wong, et al., 1999) , and by maximizing the data rate under a given power constraint in (W. Rhee and J. M. Cioffi, 2002). These are both nonlinear optimization problems with integer variables and are referred to as “marginal adaptive” (MA) and “rate adaptive” (RA) optimizations in (T. Starr et al., 1999). However, solving these problems is extremely difficult; therefore, they are only solved by relaxing the integer variables requirements and allowing real numbers. Consequently, even though this approach requires an intensive computation, it cannot yield an optimal solution. To reduce the computational complexity, some suboptimal algorithms have been proposed for MA and RA optimizations in (C. Y. Wonget al., 1999)–(Y. J. Zhang and K. B. Letaief, 2004) . Even though, the use of these algorithms still rather limited for the following reasons: In (C. Y. Wonget al., 1999), it is assumed that the average signal-to-noise ratios (SNRs) for all users are identical; while in (Y. J. Zhang and K. B. Letaief, 2004) , an equally distributed transmission power among all subcarriers is used.