A Hybrid Artifıcial Bee Colony Algorithm for the Travelling Salesman Problem


Çelikbilek Y.

3rd International Congress on Digital Business, Management, and Economics (ICDBME2023), Mersin, Türkiye, 8 - 09 Eylül 2023, (Özet Bildiri)

  • Yayın Türü: Bildiri / Özet Bildiri
  • Basıldığı Şehir: Mersin
  • Basıldığı Ülke: Türkiye
  • İstanbul Üniversitesi Adresli: Evet

Özet

Travelling Salesman Problem is one of the most famous combinatorial optimization problems dealing with minimizing a cost function or maximizing a benefit function including many independent variables. Finding the shortest path among N cities to visit each of them once is the general aim and structure of the problem. The objective function can be the calculation of the total distance of the tour or another function related with the distances among the cities in the problem. It is impossible to solve the problem in a reasonable time and way with classical optimization methods. Because the solution space contains infinite combinations and local minimums to which the solution will fit and cannot be improved. These local minimums seem like be the best solution can be much worse than the global minimums of these problems. Therefore, alternative solutions as heuristic methods or heuristic algorithms not to find the best solution but to improve the known best results have been developed by using the various algorithms and information technologies. Another important point for these algorithms is that they should reach the solution within reasonable time to have an advantage and be competitive among other methods. A hybrid Artificial Bee Colony algorithm with the operators of genetic algorithms and taboo search is proposed in this study to improve the solutions and shorten the time. Solutions obtained by running the proposed hybrid algorithm are in competition within time and range according to the best-known results.