It was supervised by Asst. Prof. Dr. Haldun Süral and Prof. Dr. Nur Evin Özdemirel. If you want to reach its full pdf, please send me an email. Here is the abstract:
In traveling salesman problem (TSP), salesman visits N cities exactly ones and returns back to where he starts. When problem size increase, it gets harder to find the optimal solution in a reasonable time; many heuristics are developed in literature for this reason. In this study, an approach that combines evolutionary algorithms, which simulate the natural evolution, with conventional heuristics is proposed. Conventional heuristics, namely, nearest neighbor, greedy and insertion, are used as crossover operators and they are restricted to a union graph generated by the parent tours, with the aim of preserving edges that are the building blocks for TSP. Performances of the algorithms are tested on problems, from the literature, with different sizes ranging from 52 to 3038. Promising results are obtained in terms of solution quality and computation time.
No comments:
Post a Comment