基于差分进化的IRA码度分布优化  

Optimization of degree distribution of Irregular Repeat-Accumulate codes based on differential evolution

作　　者：马琦[1] 王海滨[1,2] 陈曾平[1] 

机构地区：[1]国防科技大学ATR实验室,湖南长沙410073 [2]海军飞行学院教研部,辽宁葫芦岛125001

出　　处：《电子技术应用》2008年第10期116-119,共4页Application of Electronic Technique

基　　金：总装备部武器装备创新基金项目(6130524;7130524)

摘　　要：针对一般LDPC码优化方法无法有效实现IRA码度分布优化的问题,提出了特定约束下IRA码度分布的优化方法。结合密度进化的高斯近似算法优化IRA码度分布序列,提取了IRA码检验矩阵构造的特定约束以改进差分进化算法。仿真结果表明,所设计的度分布序列的噪声门限高且搜索时间比改进前减少30%。The check matrix of IRA code is different to the normal LDPC code and LDPC codes＇ design technique can not be used to design IRA codes. The method of ,optimization of irregular repeat-accumulate codes＇ degree distribution is discussed which is based on differential evolution with particular restrictions. The principle of density evolution of IRA codes is explained. Gaussian approximation to program the evolution proceeding and differential evolution are analysed, then the particular restrictions for degree distribution and improved differential evolution are discussed. Simulation results show that noise threshold of distribution sequences is high and searching time can be reduced by 30 percent.

关 键 词：低密度奇偶校验码 度分布 非规则重复累积码 差分进化 

分 类 号：TN911.72[电子电信—通信与信息系统] N941.4[电子电信—信息与通信工程]