A Novel Three-Term Conjugate Gradient Method for Efficiently Solving Unconstrained Nonlinear Optimization Problems

A new CG method is introduced in this study to solve unconstrained optimization problems, with better convergence and stability than existing approaches. The approach introduces a new update rule that greatly improves the direction of search and brings down the number of required iterations. To test if the new CG system works better than the Liu–Storey approach, a group of standard tests are performed. Fortran programming was used to compute the number of iterations and functions as presented in Table 2 and Table 3, Matplotlib on Python language and ggplot2 on R programming were used to show the results in charts.