Decoding Algorithms for LDPC Codes based on Convex Optimization and Its Applications

在本项目中，我们开发了一类基于凸优化的LDPC码译码算法。作为基于凸优化译码的其他应用，设计了一种适用于码分多址系统的多用户检测算法和一种二维均衡算法。此外，利用连续优化的概念，如梯度下降法和线性规划，提出了一类新的比特翻转算法，称为GDBF算法。

In this project, we have developed a class of decoding algorithms for LDPC codes based on convex optimization. As other applications of convex optimization based decoding, a multi-user detection algorithm for CDMA systems and a two dimensional equalization algorithm were devised. Furthermore, thorough the concept of continuous optimization, such as gradient descent method and linear programming, a novel class of bit-flipping algorithms, called GDBF algorithms, was invented.