An Adaptive Ant Colony System Memorizing Better Solutions (aACS-MBS) For Traveling Salesman Problem
Choosing the optimal among the many alternatives that meet the criteria is one of the problems that occupy life. This kind of problems frequently encountered by commercial companies in daily life is one of the issues that operators focus on with care. Many techniques have been developed that can pro...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Sakarya University
2021-06-01
|
| Series: | Sakarya Üniversitesi Fen Bilimleri Enstitüsü Dergisi |
| Subjects: | |
| Online Access: | https://dergipark.org.tr/tr/download/article-file/1384950 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Choosing the optimal among the many alternatives that meet the criteria is one of the problems that occupy life. This kind of problems frequently encountered by commercial companies in daily life is one of the issues that operators focus on with care. Many techniques have been developed that can provide acceptable solutions in a reasonable time. However, one of the biggest problems for these techniques is that the appropriate values can be assigned to the algorithm parameters. Because one of the most important issues determining algorithm performance is the values to be assigned to its parameters. The Ant Colony System (ACS) is a metaheuristic method that produces successful solutions, especially in combinatorial optimization problems. However, it is very difficult to be able to direct the algorithm to different areas of the search space and, on the other hand, to maintain its local search capability. In this study, a solution proposal is presented that updates the q0 parameter dynamically, which balances the exploitation and exploration activities of the ACS. The method has been tested on the traveling salesman problem (TSP) of different sizes, and the obtained results are evaluated together with the change in the q0 parameter, and the solution search strategy of the algorithm is analyzed. With the pheromone maps formed as a result of the search, the effect of transfer functions was evaluated. Results obtained with aACS-MBS were compared with different ant colony optimization (ACO) algorithms. The aACS-MBS fell behind the most successful solution found in the literature, by up to 4%, in large TSP benchmarks. As a result, it has been seen that the method can be successfully applied to combinatorial optimization problems. |
|---|---|
| ISSN: | 2147-835X |