A Structured Fletcher-Revees Spectral Conjugate Gradient Method for Unconstrained Optimization with Application in Robotic Model

In order to address the numerical performance issue associated with Fletcher and Reeves conjugate gradient method, a variation of spectral conjugate gradient method is presented in this paper. The spectral parameter is obtained in such a way that any line search rule is not necessary for the search direction to be sufficiently descent. The proposed scheme is globally convergent under some suitable conditions. When compared to several conventional conjugate gradient methods including CG_Descent, the preliminary numerical experiments on some set of test functions demonstrate the usefulness of the suggested method. Additionally, the effectiveness of the method is further illustrated by its success in solving robotic problems.