An Enhanced Genetic Algorithm Based Decoder for Linear Codes

Abstract

This paper introduces a new decoder based on genetic algorithms and neural networks for binary linear codes. The search space, in contrast to our previous algorithms which was limited to the codeword space, now covers the whole binary vector space.The neural network is used to favor feasible solution namely codewords. Previous genetic algorithm based decoders [2] require a lot of computing resources when used with large codes. The new decoder eludes a great number of coding operations by using the neural network. This reduces greatly the complexity of the decoder while maintaining comparable performances.