Método do gradiente conjugado Dai-Yuan modificado para otimização irrestrita
שמור ב:
| מחבר ראשי: | |
|---|---|
| Publication Date: | 2019 |
| פורמט: | Master thesis |
| שפה: | por |
| Source: | Repositório Institucional da UFG |
| Download full: | http://repositorio.bc.ufg.br/tede/handle/tede/9370 |
סיכום: | Nonlinear conjugate gradient methods are efficient first-order algorithms for solving unconstrained optimization problems. In particular, the Dai-Yuan (DY) method, introduced in the 1990s, is one of the most popular. The objective of the present work is to investigate the numerical performance of the DY method with a modification in the conjugate parameter. At each iteration of this method, it is necessary to compute a step size satisfying the standard Wolfe conditions. Thus, we will describe the line search algorithm of Moré and Thuente. Under usual assumptions, the global convergence of the modified DY method will be provided. Numerical tests will be presented using the CUTEst problem library. |