A feasible fletcher-reeves method to linear equality constrained optimization problem

In this paper, the conjugate gradient method for solving unconstrained optimization problem is applied to linear equality constrained optimization problem. An attractive property of the proposed method is that the generated direction is always feasible and descent direction, and this property is independent of the line search used. Under mild conditions, we prove that the method with Armijo-type line search is globally convergent. Numerical experiments are also given which show the efficient of the method.