Rigorous Performance Analysis of State-of-the-Art TSP Heuristic Solvers

  • Paul McMenemyEmail author
  • Nadarajen Veerapen
  • Jason Adair
  • Gabriela Ochoa
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11452)


Understanding why some problems are better solved by one algorithm rather than another is still an open problem, and the symmetric Travelling Salesperson Problem (TSP) is no exception. We apply three state-of-the-art heuristic solvers to a large set of TSP instances of varying structure and size, identifying which heuristics solve specific instances to optimality faster than others. The first two solvers considered are variants of the multi-trial Helsgaun’s Lin-Kernighan Heuristic (a form of iterated local search), with each utilising a different form of Partition Crossover; the third solver is a genetic algorithm (GA) using Edge Assembly Crossover. Our results show that the GA with Edge Assembly Crossover is the best solver, shown to significantly outperform the other algorithms in 73% of the instances analysed. A comprehensive set of features for all instances is also extracted, and decision trees are used to identify main features which could best inform algorithm selection. The most prominent features identified a high proportion of instances where the GA with Edge Assembly Crossover performed significantly better when solving to optimality.


TSP Algorithm selection EAX GPX Performance analysis 



This work was supported by the Leverhulme Trust [award number RPG-2015-395] and by the UK’s Engineering and Physical Sciences Research Council [grant number EP/J017515/1]. Results were obtained using the EPSRC-funded ARCHIE-WeSt High Performance Computer (, EPSRC grant EP/K000586/1).

Data Access. All data generated for this research are openly available from the Stirling Online Repository for Research Data (


  1. 1.
    Applegate, D.L., Bixby, R.E., Chvátal, V., Cook, W.J.: The Traveling Salesman Problem: A Computational Study. Princeton University Press, Princeton (2007)zbMATHGoogle Scholar
  2. 2.
    Helsgaun, K.: Effective implementation of the Lin-Kernighan traveling salesman heuristic. Eur. J. Oper. Res. 126(1), 106–130 (2000)MathSciNetCrossRefGoogle Scholar
  3. 3.
    Helsgaun, K.: An effective implementation of k-opt moves for the Lin-Kernighan TSP heuristic. Technical report 109, Roskilde University (2007)Google Scholar
  4. 4.
    Nagata, Y., Kobayashi, S.: A powerful genetic algorithm using edge assembly crossover for the traveling salesman problem. INFORMS J. Comput. 25(2), 346–363 (2013)MathSciNetCrossRefGoogle Scholar
  5. 5.
    Whitley, D., Hains, D., Howe, A.: A hybrid genetic algorithm for the traveling salesman problem using generalized partition crossover. In: Schaefer, R., Cotta, C., Kołodziej, J., Rudolph, G. (eds.) PPSN 2010. LNCS, vol. 6238, pp. 566–575. Springer, Heidelberg (2010). Scholar
  6. 6.
    Corne, D.W., Reynolds, A.P.: Optimisation and generalisation: footprints in instance space. In: Schaefer, R., Cotta, C., Kołodziej, J., Rudolph, G. (eds.) PPSN 2010. LNCS, vol. 6238, pp. 22–31. Springer, Heidelberg (2010). Scholar
  7. 7.
    Smith-Miles, K., Lopes, L.: Measuring instance difficulty for combinatorial optimization problems. Comput. Oper. Res. 39(5), 875–889 (2012)MathSciNetCrossRefGoogle Scholar
  8. 8.
    Pihera, J., Musliu, N.: Application of machine learning to algorithm selection for TSP. In: IEEE 26th International Conference on Tools with Artificial Intelligence, pp. 47–54 (2014)Google Scholar
  9. 9.
    Kotthoff, L., Kerschke, P., Hoos, H., Trautmann, H.: Improving the state of the art in inexact TSP solving using per-instance algorithm selection. In: Dhaenens, C., Jourdan, L., Marmion, M.-E. (eds.) LION 2015. LNCS, vol. 8994, pp. 202–217. Springer, Cham (2015). Scholar
  10. 10.
    Kerschke, P., Kotthoff, L., Bossek, J., Hoos, H.H.: Leveraging TSP solver complementarity through machine learning. Evol. Comput. 26(4), 597–620 (2018)CrossRefGoogle Scholar
  11. 11.
    Bossek, J.: netgen: Network Generator for Combinatorial Graph Problems (2016)., R package version 1.3
  12. 12.
    R Core Team: R: A language and environment for statistical computing, Vienna (2018).
  13. 13.
    Mersmann, O., Bischl, B., Trautmann, H., Wagner, M., Bossek, J., Neumann, F.: A novel feature-based approach to characterize algorithm performance for the traveling salesperson problem. Ann. Math. Artif. Intell. 69(2), 151–182 (2013)MathSciNetCrossRefGoogle Scholar
  14. 14.
    Reinelt, G.: TSPLIB - a traveling salesman problem library. ORSA J. Comput. 3(4), 376–384 (1991)CrossRefGoogle Scholar
  15. 15.
    Rohe, A.: VLSI data sets (2017). Accessed 5 Nov 2018
  16. 16.
    Tinós, R., Helsgaun, K., Whitley, D.: Efficient recombination in the Lin-Kernighan-Helsgaun traveling salesman heuristic. In: Auger, A., Fonseca, C.M., Lourenço, N., Machado, P., Paquete, L., Whitley, D. (eds.) PPSN 2018. LNCS, vol. 11101, pp. 95–107. Springer, Cham (2018). Scholar
  17. 17.
    Lin, S., Kernighan, B.W.: An effective heuristic algorithm for the traveling-salesman problem. Oper. Res. 21, 498–516 (1973)MathSciNetCrossRefGoogle Scholar
  18. 18.
    Nagata, Y., Kobayashi, S.: Edge assembly crossover: a high-power genetic algorithm for the travelling salesman problem. In: ICGA, pp. 450–457 (1997)Google Scholar
  19. 19.
    Applegate, D., Bixby, R., Chvátal, V., Cook, W.: Concorde TSP solver (2003).
  20. 20.
    Wolpert, D.H., Macready, W.G.: No free lunch theorems for optimization. IEEE Trans. Evol. Comput. 1(1), 67–82 (1997)CrossRefGoogle Scholar
  21. 21.
    Bischl, B., Mersmann, O., Trautmann, H., Preuß, M.: Algorithm selection based on exploratory landscape analysis and cost-sensitive learning, p. 313 (2012)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Computing Science and MathematicsUniversity of StirlingStirlingUK
  2. 2.Univ. Lille, CNRS, Centrale Lille, UMR 9189 - CRIStALLilleFrance

Personalised recommendations