| Summary: | This paper studies a new hybrid conjugate gradient (CG) method based on the Aini-Rivaie-Mustafa (ARM) CG method for solving nonlinear unconstrained optimization problems. The new hybrid method eliminates the negative values generated by the ARM method in its CG coefficient by replacing those values with a positive CG coefficient. The numerical test was conducted on 10 standard test functions from small to large scale under inexact line search. Based on the numerical results, the method proved to be more efficient compared with some older versions of CG method in terms of number of iteration and CPU time. In addition, a set of data for number of road accidents was collected from Portal Rasmi Polis Diraja Malaysia. By using discrete least squares method of numerical analysis and CG method in unconstrained optimization, the data can be estimated. Results from the error calculation for both methods showed that the proposed CG method is comparable with the least squares method. © 2024 AIP Publishing LLC.
|
|---|