This code implements an unconstrained optimization algorithm based on a non-monotone truncated Newton Method.
At each iteration a truncated newton method, based on the conjugate gradient algorithm, is used to find the search direction; the step size along the search direction is computed using a nonmonotone generalization of the Armijo's line search method. This algorithm is a detailed implementation of the work presented in "A Truncated Newton Method with Nonmonotone Line Search for Unconstrained Optimization", L. Grippo, F. Lampariello, AND S. Lucidi.
For more information on Non-Monotone Line search algorithms visit my home page at http://www.marioantonelli.it/.