Algoritimos Heurísticos Aplicados à detecção Multiusuário DS/CDMA

Abstract

The characteristics of nine heuristic algorithms based on genetic evolution theory and local search, applied to a DS/CDMA (Direct Sequence/Code Division Multiple Access) multi-user detection problem in Flat Rayleigh fading channel are introduced. The algorithms comparisons through the performance versus computational complexity tradeoff are carried out. The computational complexity is determined in terms of the number of operations to reach the Maximum Likelihood (ML) performance. The effects on the performance of the errors in the estimates of the channel coefficients are also considered, as well as the performance degradation due to overload. There are three contributions in this work: a) a compared analysis of the evolutionary, evolutionary with cloning and local search algorithms applied to MuD detector in Flat Rayleigh channels; b) a systematic comparison of these algorithms in terms of the same simulation scenario and performance analysis base; c) a more fair and precise complexity analysis for the MuD detector in terms of the number of the computational operations.