A modified of Dai and Yuan method for solving unconstrained optimization problems

The Conjugate Gradient (CG) method is a widely used and efficient approach for solving large-scale unconstrained optimization problems due to its simplicity and low memory demands. Despite its success, ongoing research seeks to enhance its performance and robustness. This paper proposes a novel parameter modification to the classical Dai-Yuan CG formula, aiming to improve numerical behavior, reduce iteration counts, and boost convergence speed and stability. Theoretical analysis confirms that the new method ensures strong global convergence and descent properties.