利用遗传算法构造QC-LDPC码  

Construction of QC-LDPC Codes with Genetic Algorithm

作　　者：郑丹玲[1] 穆攀[1] 田凯[1] 袁建国[1] 

机构地区：[1]重庆邮电大学光通信与网络重点实验室,重庆400065

出　　处：《电讯技术》2015年第4期355-359,共5页Telecommunication Engineering

基　　金：国家自然科学基金资助项目(61171158);重庆市自然科学基金资助项目(cstc2013jcyjA40052;cstc2012jjA40060);重庆市教委科学技术研究项目(KJ130515)~~

摘　　要：考虑到围长(girth)对低密度奇偶校验(LDPC)码的影响,提出了一种利用遗传算法构造大girth的准循环LDPC(QC-LDPC)码的新方法。该方法借助于计算机搜索,多次运用遗传算法,分步提高girth,在得到大girth的同时,构造出具有准循环结构的LDPC码。分析发现,该构造方法的复杂度与码长成线性关系。仿真结果表明:在误码率(BER)为10-6时,新方法构造的QC-LDPC码比基于欧式几何构造方法、Gallager和Mackay构造法分别获得约0.15dB、0.5dB和0.2dB的净编码增益(NCG),且因具有准循环结构更易于存储和硬件实现。A new method is proposed to construct a large girth quasi-cyclic low density parity check（ QC- LDPC） code with Genetic Algorithm（GA） by consideration of LDPC codes under the influence of girth. This method depends on computer search, uses GA repeatedly,improves girth step by step. A large girth is obtained,at the same time LDPC codes with a quasi-cyclic structure is constructed. Analysis shows its complexity has a linear relationship with code length. Simulation results illustrate that when the bit error rate（BER） is 10^-6 QC-LDPC codes constructed with the new method has net coding gain（NCG） of 0.15 dB,0.5 dB,0.2 dB over LDPC code based on Euclidean Geometry, Gallager random codes and Mackay random codes, respectively, and it is easy to restore and be implemented in hardware because of quasi-cy- clic structure.

关 键 词：QC-LDPC 大围长 遗传算法 复杂度分析 

分 类 号：TN911.22[电子电信—通信与信息系统]