Analysis of Human Performance in the Solution of Traveling Salesman Problem

Kenan KARAGUL1, Yusuf SAHIN2,*, Necdet GUNER3, Aykut ORAL4
1Pamukkale University, TURKEY
2Burdur Mehmet Akif Ersoy University, TURKEY
3Pamukkale University, TURKEY
4Pamukkale University, TURKEY
DOI: 10.14689/ejer.2020.87.7


Purpose: Traveling Salesman Problem (TSP) that can be extended and modified in various ways, is a practical and realistic type of problem and forms the basis for the visual and spatial solution of many optimization problems. In this study, 15, 25 and 35 nodes Travelling Salesman Problems were solved by secondary school, high school and undergraduate students in order to examine human performance in the solution of TSP. In addition to this assessment, whether gender and education level had an impact on the quality of the solution was analyzed.

Research Methods: The categorical comparisons of solutions, male-female, and educational level were examined with the help of nonparametric statistical methods. In addition, for the three levels of education, the Kruskal-Wallis Test was applied to determine whether the difference between the education levels was significant. On the other hand, the performance of human solutions was compared with the heuristic methods found in the literature.

Findings: As a result, it was seen that the gender difference was not statistically significant for all problems. On the other hand, it was determined that the education level had a significant effect on the solutions. It can be concluded for the given problems that the human solutions produced as good results as the solutions obtained with other heuristic methods in the literature.

Implications for Research and Practice: The findings of the study confirm previous studies. By examining the effect of the factors that affect optimization strategies, it is possible to produce human-based TSP heuristic solutions that surpass all existing heuristic algorithms.

Keywords: Problem-solving, optimization, cognitive researches.