An Analysis of Two Genetic Algorithm Parameters in Two Optimization Problems
Philippine Journals Online
View Archive Info| Field | Value | |
| Title |
An Analysis of Two Genetic Algorithm Parameters in Two Optimization Problems
|
|
| Creator |
Proceso L. Fernandez, Jr.; Department of Information Systems and Computer Science, Ateneo de Manila University
|
|
| Subject |
Science; Engineering; Information System; Computer Science
Genetic Algorithm; evolutionary computation |
|
| Description |
A Genetic Algorithm (GA) has many parameters that can be modified to improve the performance of particular implementations. These parameters include the type of selection, crossover point, mutation probability, replacement policy, and terminating condition. In this paper, a GA was used to solve two optimization problems involving single-variable functions that are differentiable at every point in their domains. Two GA parameters, i.e., the tournament size for selection and the probability of mutation, were modified. Statistical information such as average number of generations before obtaining the optimal results and the average (best) fitness value after a fixed number of generations were obtained in order to analyze how the varying values of these parameters affect the performance of the particular GA implementation. This study showed that the values used for these two parameters provided extremely good results for one problem but generated sub-optimal results when applied to a different problem.
|
|
| Publisher |
Loyola Schools, Ateneo de Manila University
|
|
| Contributor |
—
|
|
| Date |
2008-05-22
|
|
| Type |
—
|
|
| Format |
application/pdf
|
|
| Identifier |
http://www.philjol.info/index.php/LSR/article/view/290
|
|
| Source |
Loyola Schools Review; Vol 6 (2007); 78-95
|
|
| Language |
en
|
|
| Coverage |
Philippines
— — |
|
| Rights |
The individual authors assert the copyright of their works, and the Office of Research and Publications asserts the copyright of the journals under the Intellectual Property Code of the Philippines [Republic Act No. 8293] and any provisions thereof. For inquiries on multiple reproduction or publication of the works, please contact orp@admu.edu.ph
|
|