中文摘要：

为了解决多进制低密度奇偶校验（LDPC）码的通用编码，从Tanner图结构出发，利用下三角和上三角（LU）分解进行编码的算法，以保证矩阵稀疏性为目标，详细推导了与分析行主元策略、行列主元策略和行列相乘主元策略等主元选取策略，并对所提算法进行了仿真．测试结果表明，相比于现有LDPC码LU分解编码方法，新算法能将矩阵稠密度降低一半以上，为多进制LDPC码通用编码算法的直用奠定了基础．

英文摘要：

For general encoding algorithm of low density parity check （LDPC） codes over GF （q） , a Tanner graph and research encoding algorithm is presented based on lower upper （LU） decomposition. In order to keep sparsity of matrix, three pivot selection criterias named row pivot, row-column pivot and product of row-column pivot are deduced and analyzed. Simulations show that the new algorithm could be lower than one half density compared with existing algorithm. The conclusion is useful to application of universal encoding algorithm for LDPC codes over GF（q）.