The employment of multiple antennas in Orthogonal Frequency Division Multiplexing (OFDM) based systems constitutes an effective way of improving the systems' performance. Multiple transmitters may be used for achieving a number of design objectives. This thesis explores two different design principles. Firstly, the family of Space-Time Block Codes (STBCs) can be exploited, which exhibits a remarkable encoding and decoding simplicity and yet achieves a high transmit diversity gain, when combined with a range of Forward Error Correction (FEC) codes, such as Turbo Convolutional (TC) codes, Low Density Parity Check (LDPC) codes and Coded Modulation (CM) schemes. Secondly, multiple antennas may be used for supporting a multiplicity of users by differentiating them with the aid of their unique, user-specific Channel Impulse Responses (CIRs). More explicitly, by invoking the Multiple-Input Multiple-Output (MIMO) Space Division Multiple Access (SDMA) techniques, a number of simultaneous users may share the same frequency bandwidth in different geographical locations, which results in an increased bandwidth efficiency. In order to improve the achievable performance of SDMA-OFDM systems, various CM schemes, namely Trellis Coded Modulation (TCM), Turbo TCM (TTCM), Bit-Interleaved Coded Modulation (BICM) and Iteratively Decoded BICM (BICM-ID) can be employed for attaining a substantial coding gain without any bandwidth expansion. Furthermore, classic Walsh-Hadamard Transform Spreading (WHTS) often used in Code Division Multiple Access (CDMA) systems can be invoked across a number of OFDM subcarriers in the frequency domain, for the sake of exploiting the frequency diversity potential offered by frequency-selective fading channels. Regardless of the specific choice of the FEC schemes, the Multi-User Detector (MUD) invoked at the SDMA base station has a significant effect on the achievable system performance. Specifically, the Maximum-Likelihood (ML) MUD achieves the optimum performance at the cost of a typically excessive computational complexity. By contrast, the Genetic Algorithm (GA) based MUDs attain a near-ML performance, despite having a substantially reduced complexity. Moreover, the GA MUDs can be further enhanced in terms of three different design aspects. Firstly, an iterative detector architecture involving channel decoders can be employed. Secondly, the GA's detection capability can be improved by introducing a novel genetic mutation approach, which we refer to as the Biased Q-function based Mutation (BQM) scheme. Thirdly, it is beneficial to generate soft information as the GA's output, which assists the channel decoders in attaining an increased coding gain. Furthermore, time-domain Direct-Sequence Spreading (DSS) and a novel subcarrier-based Slow Frequency-Hopping (SFH) technique, referred to as Slow SubCarrier-Hopping (SSCH), can be incorporated into SDMA-OFDM systems. With the aid of the new Uniform SSCH (USSCH) patterns proposed, hybrid DSS/SSCH SDMA-OFDM systems are rendered capable of overcoming a number of disadvantages of the conventional SDMA-OFDM arrangements, benefiting from both a high frequency diversity gain and a robustness against the Multi-User Interference (MUI).