Enhancing Convergence Speed of Standard Ant Colony System

Creative Commons License

Sarıkoç F.

3rd International Conference on Engineering and Applied Natural Sciences, Konya, Türkiye, 14 - 17 Ocak 2023, ss.200-210

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Basıldığı Şehir: Konya
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.200-210
  • Erciyes Üniversitesi Adresli: Evet

Özet

In this paper, we focus on a hybrid implementation of the Ant Colony System (ACS) algorithm for the traveling salesman problem. The motivation for this hybrid approach is to combine the divergence capabilities of ACS with the fast convergence capabilities of the Tabu Search (TS) algorithm. To achieve this aim, we modified the standard ACS algorithm to improve convergence speed with a frequency parameter inspired by the TS algorithm. We added a new parameter to reinforce the exploration more around the best global solution and experimented with standard ACS and the proposed ACS approaches on some TSP problems. As a result of the experiments, we observed faster convergence performance in the modified ACS algorithm.