This dissertation explores the application of Genetic Algorithms (GAs) assisted multiuser detection in the context of Code Division Multiple Access (CDMA). The optimum multiuser detector proposed by Verdu [1] entails searching for a particular K-bit sequence that optimises the correlation metric, where K is the number of users. Hence, it has a computational complexity that is exponentially proportional to the number of users and its implementation becomes impractical, when there is a high number of users. GAs have been successfully applied to solve complex optimisation problems in many fields. Hence in this dissertation, we will investigate the feasibility of employing GAs in solving the optimum multiuser detection problem. We commence by determining a set of GA configurations that are capable of offering a near-optimum performance at the cost of a reduced computational complexity, compared to the optimum multiuser detector receiving over a simple AWGN channel. Our study showed that certain GA parameters substantially influence the overall performance of the detector. More importantly, we will show that the optimum performance can be achieved up to a certain SNR value at a complexity less than half of that required by the optimum multiuser detector. The employment of the GA-assisted multiuser detector is then extended to an asynchronous CDMA system. Antenna diversity based on the Pareto optimality approach can further improve the achievable performance. The proposed GA-assisted multiuser detector is then extended further, so that Channel Impulse Response (CIR) estimation can also be performed jointly by the same GA without incurring any additional computational complexity and without requiring training symbols. Hence the joint GA-assisted channel estimator and symbol detector is capable of offering a higher throughput and a shorter detection delay, than that of explicitly trained CDMA multiuser detectors.