GA-based soft-decision decoding of block codes

Abstract

In this paper, soft-decision decoding is considered as an optimisation problem and a computationally efficient soft-decision decoding algorithm for block codes is developed using a compact Genetic Algorithm. The performance of this algorithm is investigated and compared with other decoding schemes. The results show the proposed algorithm gives large gains over the Generalised Minimum Distance (GMD) decoding algorithm and algebraic hard-decision decoding in flat Rayleigh fading. Further, the proposed algorithm achieves near maximum likelihood decoder (MLD) performance for the codes simulated. The complexity of this algorithm is also discussed and compared to the standard Order-l reprocessing algorithm.