Implementation of golden section search method in SAGE algorithm

The SAGE algorithm gives the chance to divide the multi-dimension optimization problem, which one confronts when the maximum likelihood (ML) estimation is performed, into iterations of one-dimension optimization problems. The progress to find out the maximum of the log-likelihood function during Maximization-step (M-step) of the algorithm is yet to be refined with purpose of reducing the time cost of the computation and also path management for channel parameter estimation. In this contribution, we try to improve the performance of the SAGE algorithm by using Golden Section search method in the M-step. Numerical and experimental results demonstrate that the effectiveness of the SAGE algorithm is improved.