New iterative decoding algorithms of 2D product block codes

This paper presents four new iterative decoders for two dimensional product block codes (2D-PBC) based on Genetic Algorithms. Each of these iterative decoders runs in parallel on a number of processors connected by a network. They have almost the same complexity as the conventional iterative decoder, but their performances are improved since at each iteration, they trap the better of extrinsic information computed by the elementary decoders running simultaneously on all processors.