摘要
提出一种基于内点法的改进重构算法,尝试用专门的内点算法解决稀疏重构问题。首先,在内点法基础上引入预处理算子重新设计来避免牛顿方程系统的构造,使矩阵拥有良好的可调性;其次,利用稀疏矩阵的矩阵特性简化矩阵矢量增量。仿真实验结果表明改进的内点算法对实际问题的处理是有效且优于其他算法的。
To use a special interior point method solve the sparse reconstruction problem,a modified reconstructed method based on interior point method was proposed. First,to avoid construction of Newton equations system,a preconditioning operator was lead up which was based on the interior point method to make the matrix have good adjustability. Second,the characteristics of sparse matrix were used to simplify the matrix-vector incrementing. The experimental results show that the modified interior point method is effective and better than any other algorithms for practical problems.
出处
《仪表技术与传感器》
CSCD
北大核心
2015年第6期138-142,共5页
Instrument Technique and Sensor
基金
国家科技重大专项资助项目(2009ZX02308-004)
关键词
压缩感知
内点法
稀疏矩阵
预处理
共轭梯度
compressed sensing
interior point method
sparse matrix
precondition
conjugate gradient