An efficient soft decision decoding algorithm using cyclic permutations and compact genetic algorithm

The compact genetic algorithm cGA is used in this paper to design an efficient soft-decision decoding algorithm, especially for the cyclic codes, because the cGA dramatically reduces the population's size and rapidly converges to the optimal solution compared to classical genetic algorithms. Our main contribution is to exploit the cyclic property of cyclic linear codes to reduce the complexity of the decoding process especially in the test sequences generation and re-encoding stage where we use the generator polynomial instead of the generator matrix. The second idea behind our decoding algorithm is the complexity improvement inside of cGA by decreasing the probability vector's length, which becomes less than the length of the cGA original one. The experiments were carried out on the most popular cyclic codes, and the results show that the performances of our algorithm are better than some famous decoding algorithms in terms of Bit Error Rate.