AUTHOR=Mariescu-Istodor Radu , Fränti Pasi 

TITLE=Solving the Large-Scale TSP Problem in 1 h: Santa Claus Challenge 2020

JOURNAL=Frontiers in Robotics and AI

VOLUME=Volume 8 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/robotics-and-ai/articles/10.3389/frobt.2021.689908

DOI=10.3389/frobt.2021.689908

ISSN=2296-9144

ABSTRACT=Scalability of TSP algorithms to handle large-scale problem instances has been an open problem for long time. We arranged a so-called Santa Claus challenge and invited people to submit their algorithms to solve a TSP problem instance larger than 1M nodes given only 1 hour of computing time. In this paper, we analyze the results and show which design choices are decisive to provides the best solution to the problem with the given constraints. Results imply that local search with k-opt operator up to k=5 is still the state-of-the-art. The most important design choice turned out to be the localization of the operator using neighborhood graph. Divide-and-merge strategy suffered a 2% loss of quality, however, with powerful architecture and through parallelization such a result can be obtained in less than 2 minutes which can make the difference in some real life application.