Min-Max decoding for non binary LDPC codes

with the increase in requirements on fiber-optical communication system by people, transmission quantity has been increased. In order to guarantee quick and effective fiber-optical communication, people have to use LDPC codes to implement correction modulation, of which, the ability of non-binary LDPC codes to correct abrupt error and random error become more stronger, and it is corresponds with high-order modulation, it is suitable to be used in optical transmission system with super speed and long distance, so it has become to the key point researched by people. This paper starts from non-binary LDPC codes to analyze the principle application of correcting matrix and Tanner diagram, it also illustrates coding and decoding of LDCP codes and puts forward Min-Max non binary LDPC codes algorithm, which simplifies calculation of non binary LDPC codes and can effectively promote enhancement of communication in error correction. Introduction As the soft decision technology universally researched by people, LDPC has excellent ability in error correction; error level is low and can be concurrently realized. Meanwhile, research on binary LDPC codes is relatively mature, including binary LDPC codes and decoding algorithm, capacity analysis method, search algorithm of optimal degree based on Gaussian white noise channel and Rayleigh fading channel with irregular code. However, compare non-binary LDPC code with LDPC codes, its ability in abrupt error correction and random error correction is stronger, and it is more suitable with high-order modulation, so it is suitable to be used in optical transmission system with super speed and long distance. Overview on non-binary LDPC codes LDPC codes was put forward by Gallager in 1960, it is the linear block code, ix represents better capacity in data transmission and data storage, it is exclusively determined by generation matrix G or parity check matrix H, so it can define LDPC codes by parity check matrix H. Of which, H has 4 natures, which means each line has p 1, each kind has r and 1, the position between any 2 lines is the same and the number of valuing 1 dose not exceed 1, the line number comparison between p, t as well as code length and H is very small. From these natures we can see that parity check matrix and H respectively has special line weight p and parallel weight r, and any 2 lines have 1 exceed the same position, the density of 1 is very small, so it is called as parity check matrix with low density, the details are indicated by the following figure 1: Figure 1 Parity check matrix with low=density of LDPC Except to use check matrix to express LDPC codes, it can also use bidirectional graph model to 4th International Conference on Machinery, Materials and Information Technology Applications (ICMMITA 2016) Copyright © 2017, the Authors. Published by Atlantis Press. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/). Advances in Computer Science Research, volume 71