A SURVEY OF SPECTRAL CONJUGATE GRADIENT METHODS FOR NONLINEAR OPTIMIZATION PROBLEMS

Spectral Conjugate Gradient (SCG) methods are a new class of unconstrained optimization algorithms which are characterized by low memory requirements and strong local and global convergence properties. In this paper, we investigate some well-known SCG methods by focus on descent property and global convergence of them.

کلیدواژه ها:

Spectral conjugate gradient method ، Unconstrained optimization problem ، Global convergence ، Wolfe type line search