A globally convergent hybrid conjugate gradient method with strong Wolfe conditions for unconstrained optimization