A modified Polak–Ribi‘ere–Polyak descent method for unconstrained optimization

In this paper, a modified Polak–Ribi‘ere–Polyak (MPRP) conjugate gradient method for smooth unconstrained optimization is proposed. This method produces at each iteration a descent direction, and this property is independent of the line search adopted. Under standard assumptions, we prove that the MPRP method using strong Wolfe conditions is globally convergent. The results of computational experiments are reported and show the effectiveness of the proposed method.