A new FA property based genetic algorithm for improving source and channel estimates

We introduce a new genetic algorithm, which exploits the finite alphabet (FA) property to refine the estimates of information source symbols and channel estimates obtained by any identification algorithm. The new genetic source symbol refinement algorithm (GSR) is tested to cope with rapidly time-varying finite impulse response (FIR) channels with additive noise model. Computational results show that, as compared with recent sophisticated alternatives for the problem, it offers superior performance with reduced complexity.