Decoding of Block Codes by using Genetic Algorithms and Permutations Set

Abstract

Recently Genetic algorithms are successfully used for decoding some classes of error correcting codes. For decoding a linear block code C, these genetic algorithms computes a permutation p of the code generator matrix depending of the received word. Our main contribution in this paper is to choose the permutation p from the automorphism group of C. This choice allows reducing the complexity of re-encoding in the decoding steps when C is cyclic and also to generalize the proposed genetic decoding algorithm for binary nonlinear block codes like the Kerdock codes. In this paper, an efficient stop criterion is proposed and it reduces considerably the decoding complexity of our algorithm. The simulation results of the proposed decoder, over the AWGN channel, show that it reaches the error correcting performances of its competitors. The study of the complexity shows that the proposed decoder is less complex than its competitors that are based also on genetic algorithms.