An efficient evolutionary algorithm for multiuser detection in CDMA systems

This paper introduces an adaptive ( μ + λ ) evolutionary algorithm (EA) for multiuser detection (MUD) in direct sequence code division multiple access (DS-CDMA) systems. The EA based multiuser detector takes the output of parallel interference cancellation (PIC) detector as the initial population and adopts maximum likelihood decision rule to detect the user bit sequences. During an EA running, the offspring population size is adaptively adjusted according to the current success probability. The major advantage of the adaptive EA based multiuser detection scheme is that it obtains rather good bit error rate (BER) performance with lower computational complexity. Monte Carlo simulation results show the EA based multiuser detector can always converge to the optimal solution with a small number of generations. Numerical analysis shows its computational time is polynomial complexity and less than other heuristic search algorithms.