New Three-Term Conjugate Gradient Method Based on WFR Variant with Practical Applications

Abstract

The conjugate gradient method is widely used for large-scale unconstrained optimization due to its efficiency in iteration number and computing time. In this dissertation, we propose a new three-term conjugate gradient method inspired from the Fletcher–Reeves approach, where we introduce a novel descent direction. We conduct the numerical experiments on classical unconstrained optimization problems and on practical applications including the image restoration and sparse signal recovery problems. The results confirm the efficiency and robustness of the proposed approach compared to some existing algorithms